[marluc]

Hallo,

I am solving a three dimensional pipe flow with a non-uniform applied gradient on half its surface. I am using the following groovyBC:

     wallHeat
     {
         type                    groovyBC;
         gradientExpression      "2e5/4*pi*cosTheta";
         fractionExpression      "0";
         variables               "r=sqrt(sqr(pos().x)+sqr(pos().y));cosTheta=mag(pos().x/r);";
         refGradient             uniform 50000;
         refValue                   uniform 50000;
         value                   uniform 50000;
     }

solve for a volScalarField T and from it I calculate a volScalarField T1.

The internalField values of T1 are correct but the boundary values on the above patch are not because they are always set at the specified uniform value and not as a nonUniform List as for the field T:

For T1:

 wallHeat
    {
        type            groovyBC;
        refValue        uniform 50000;
        refGradient     uniform 50000;
        valueFraction   uniform 1;
        value           uniform 50000;
        valueExpression "0";
        gradientExpression "2e5/4*pi*cosTheta";
        fractionExpression "0";
        evaluateDuringConstruction 0;
        variables       
2
(
"r=sqrt(sqr(pos().x)+sqr(pos().y));"
"cosTheta=mag(pos().x/r);"
)
;
        timelines       (
);
        lookuptables    (
);

For T:

  wallHeat
 {
     type            groovyBC;
     refValue        uniform 0;
     refGradient     nonuniform List<scalar>
400
(
157049
156807
156323
155598
.
.
.       
 valueFraction   uniform 0;
 value           nonuniform List<scalar>
 400
(
479.485
479.069
478.242
477.005
.
.
.
)
;
         valueExpression "0";
         gradientExpression "2e5/4*pi*cosTheta";
         fractionExpression "0";
         evaluateDuringConstruction 0;
         variables
 2
(
 "r=sqrt(sqr(pos().x)+sqr(pos().y));"
 "cosTheta=mag(pos().x/r);"
 )
 ;
         timelines       (
);
         lookuptables    (
 );

As you can see it seems that the non-uniform gradient BC is not evaluated for T1. But why? Is it because I don't solve any equation for that field but I just derive it from T? But if I use a fixedGradient BC then everything works finr and I obtain the correct T1 values at the wall.

Can someone help?

Thank you in advance,
Luca