[r2d2]

Hi All, a question, maybe someone can help...
I use groovyBC for some scalar, say X, and I want to change the "value" field at runtime (actually I want to eliminate some negative values out of this
value nonuniform List<scalar> etcetc...
I used to do it for members of
mixedFvPatchField<vector> &U1Patch = refCast<mixedFvPatchVectorField>(U1.boundaryField( )[labelH2out]);
..e.g.
U1Patch.refValue()[iFaceLocal] = SomeVector (0,0,umag_new);
U1Patch.valueFraction()[iFaceLocal] = 1; (from zero gradient to dirichelt)

Is this possible? Or cause groovy it´s loaded only at runtime it is not...
many thanks,
radu

[gschaider]

Originally Posted by r2d2

Hi All, a question, maybe someone can help...
I use groovyBC for some scalar, say X, and I want to change the "value" field at runtime (actually I want to eliminate some negative values out of this
value nonuniform List<scalar> etcetc...
I used to do it for members of
mixedFvPatchField<vector> &U1Patch = refCast<mixedFvPatchVectorField>(U1.boundaryField( )[labelH2out]);
..e.g.
U1Patch.refValue()[iFaceLocal] = SomeVector (0,0,umag_new);
U1Patch.valueFraction()[iFaceLocal] = 1; (from zero gradient to dirichelt)

Is this possible? Or cause groovy it´s loaded only at runtime it is not...
many thanks,
radu

Hi R2D2!

To quote another StarWars-Character "Not clear your question is, young Yedi. Changing field at run-time purpose of boundary condition is"

Where do you run the code-snipplet that you show?
- in a special BC
- in the solver
- in a special postprocessing-utility?

In groovyBC what you're trying to do can be written like this (I hope the names of the parameters are right. Just because I wrote that stuff doesn't mean that I know how to use it):

type groovyBC;
valueExpression "vector(0,0,umag_new)";
gradientExpression "vector(0,0,0)";
fractionExpression " (scalar<0) ? 1 : 0";

If you want to do this after the simulation you can use funkySetBoundaryField (part of swak4Foam) to manipulate the boundary fields (but I havn't tried this as a postprocessor for large cases. Might take a while)

To end: "If answer you not like question reformulate you must. Clear asking of question most noble power of the Yedi is"

[r2d2]

Oh groovyBC Master,
Let me explain myself...
Me have a scalarField with a groovyBC patch and during runtime
some members in "value" go negative..and I want to cut that off
(make them zero).
E.g.
variable "X" with
myPatch
{
type groovyBC;
refValue uniform 0;
refGradient uniform 0;
valueFraction uniform 1;
value nonuniform List<scalar>
10
(
0.0985793
0.0985383
0.0985488
0.0985484
0.0985443
0.09854
0.0985374
-0.00001
-0.0001
-0.001234
)
;
valueExpression "Ccref";
gradientExpression "0";
fractionExpression "f";
variables "vel=mag(U);A=-nuCceff;B=vel;C=vel;f=1.0/(1.0+A/(B*mag(delta())));Ccref=C/B;";
timelines ();
}

...and...if I do (in the code)

label Mylabel = mesh.boundaryMesh().findPatchID("myPatch");
scalarField &Xfldlabel=X.boundaryField()[Mylabel];

forAll(Xfldlabel,fI)
{
if (Xfldlabel[fI]<0.0) Xfldlabel[fI]=0.0
}
X.correctBoundaryConditions();

runtime.writeNow();
return(0);

...my "value" field in the X for the groovy BC stays the same...as in like it did not do what I asked it to do in the forAll loop...


That´s the "problem" this young Jedi is facing...any help appreciated...
Regards,
r2d2

[gschaider]

Originally Posted by r2d2

Oh groovyBC Master,
Let me explain myself...
Me have a scalarField with a groovyBC patch and during runtime
some members in "value" go negative..and I want to cut that off
(make them zero).
E.g.
variable "X" with
myPatch
{
type groovyBC;
refValue uniform 0;
refGradient uniform 0;
valueFraction uniform 1;
value nonuniform List<scalar>
10
(
0.0985793
0.0985383
0.0985488
0.0985484
0.0985443
0.09854
0.0985374
-0.00001
-0.0001
-0.001234
)
;
valueExpression "Ccref";
gradientExpression "0";
fractionExpression "f";
variables "vel=mag(U);A=-nuCceff;B=vel;C=vel;f=1.0/(1.0+A/(B*mag(delta())));Ccref=C/B;";
timelines ();
}

...and...if I do (in the code)

label Mylabel = mesh.boundaryMesh().findPatchID("myPatch");
scalarField &Xfldlabel=X.boundaryField()[Mylabel];

forAll(Xfldlabel,fI)
{
if (Xfldlabel[fI]<0.0) Xfldlabel[fI]=0.0
}
X.correctBoundaryConditions();

runtime.writeNow();
return(0);

...my "value" field in the X for the groovy BC stays the same...as in like it did not do what I asked it to do in the forAll loop...


That´s the "problem" this young Jedi is facing...any help appreciated...
Regards,
r2d2

Reordering words my habit is. The droid only says "Bip bip. Beeep. Tweeee"

Anyway. That is a bit strange. BTW: my suspicion is that you're on a wall with a zero-velocity. That's why the values of the super-class (refValue, refGradient and valueFraction evaluate to the values you have)
As they stay constant: could you try to set the type from groovyBC to mixed (with the above ref-values) and see if the resulting values are the same. If yes: then it is a problem with the superclass (mixed)

Bernhard

[r2d2]

Hi again,
Thing is that this is an exit and for my "X" I wanted to set up a bc of type
A*dX/dn+B*X=C with A,B, C functions of the velocity (sort of
convective-diffusive flux bc)..the trouble that I ran into at runtime was that X on this patch went negative-when it is a mass fraction(!?) (my posted example was a bit "forced", the values were low ..of order -1e-5) and I wanted to shave them off ...hence my question whether I could access this "value" member at and play with it with correctBoundaryConditions and so on much as I was doing the trick with a mixedFvPatch ...
My guess now is that as I load groovy via controlDict as a .so the code will understand it at runtime, but in the code itself I won´t be able to use/access members of the class...have to find a way round it.
Many thanks for your hep anyways.
r2d2

[gschaider]

Originally Posted by r2d2

Hi again,
Thing is that this is an exit and for my "X" I wanted to set up a bc of type
A*dX/dn+B*X=C with A,B, C functions of the velocity (sort of
convective-diffusive flux bc)..the trouble that I ran into at runtime was that X on this patch went negative-when it is a mass fraction(!?) (my posted example was a bit "forced", the values were low ..of order -1e-5) and I wanted to shave them off ...hence my question whether I could access this "value" member at and play with it with correctBoundaryConditions and so on much as I was doing the trick with a mixedFvPatch ...
My guess now is that as I load groovy via controlDict as a .so the code will understand it at runtime, but in the code itself I won´t be able to use/access members of the class...have to find a way round it.
Many thanks for your hep anyways.
r2d2

In groovyBC (without c++) itself this is not possible. Anyway. The interesting question for me is whether if you leave the ref-values like in your example and set "type mixed" the value-field has the same values.

Bernhard