[gfilip]

Hello,

I am experimenting with groovyBC and can successfully use the valueExpression on a patch. When I try to set a gradientExpression for pressure, such as:

   
inlet
    {
        type            groovyBC;
        //valueExpression               "100";
        gradientExpression      "100";

    }

and then calculate the normal gradient of pressure on the inlet, I get non-constant values much different than 100. When using the gradientExpression I also obtain a 0 pressure on the inlet boundary. I also tried using:
gradientExpression "vector(100,0,0)&normal()" but that gives the same result.

This is a simple 2D case with an oscillatingFixedValue velocity inlet ran using pisoFoam.

Am I interpreting the gradientExpression patch parameter wrong?

[akidess]

From the documentation:

* fractionExpression 
   Determines whether the face is Dirichlet (1) or Neumann (0). Defaults to 1

You can probably see the problem. Fix fractionExpression and all will be good.

[gfilip]

Hi Anton,

I have in fact tested the fractionExpression as well, but setting it to 0 did not help.

[gschaider]

Originally Posted by gfilip

Hello,

I am experimenting with groovyBC and can successfully use the valueExpression on a patch. When I try to set a gradientExpression for pressure, such as:

   
inlet
    {
        type            groovyBC;
        //valueExpression               "100";
        gradientExpression      "100";

    }

and then calculate the normal gradient of pressure on the inlet, I get non-constant values much different than 100. When using the gradientExpression I also obtain a 0 pressure on the inlet boundary. I also tried using:
gradientExpression "vector(100,0,0)&normal()" but that gives the same result.

This is a simple 2D case with an oscillatingFixedValue velocity inlet ran using pisoFoam.

Am I interpreting the gradientExpression patch parameter wrong?

I'm just wondering whether I'm misinterpreting your posting. "I get non-constant values much different than 100" Do you mean "I calculated the gradient using the value in the next cell and the delta coefficient and it was much different from 100". In that case there is a problem. Or did you mean "'value' on the patch is different from 100". That is to be expected because you're only setting the gradient on the patch (basically the relative change to the next cell) NOT the value with a Neuman condition

[tfuwa]

Hi All,

I would like to set up a gradient B.C., ∂p/∂x = cos (2*pi*t) .

Is there a way to do this with groovyBC or openfoam official tools?

Cheers.

[gschaider]

Originally Posted by tfuwa

Hi All,

I would like to set up a gradient B.C., ∂p/∂x = cos (2*pi*t) .

Is there a way to do this with groovyBC or openfoam official tools?

In groovyBC this would be something like this:
valueExpression "0"; // this doesn't matter
fractionExpression "0"; // no value, all gradient
gradientExpression "sin(2*pi*time())";

Official tools: you can do something similar with some C++-programming with the codedMixed-BC I guess

[tfuwa]

Thanks Bernhard! It works perfect.