[Saideep]

Hi guys;

I am dealing with the interfoam solver and modifying the standard interfoam of 2.3.1 version.

I came across a strange piece of code which i have not understood. I appreciate anyone's help here.

I calculate the capillary pressure developed from the capillary(surface) forces. I calculate them at the cell center and interpolate them to the face centre. {so conversion from volScalarField to surfaceScalarField}.
I solve for the fvScalarMatrix form equation to find the capillary pressure. Till here its looks fine. Compiles and case runs.

Now, I would like to use the gradient of the calculated capillary pressure for some filtration process.
Just checking the workability for a smaller term I created a new term:
surfaceScalarField newTerm = fvc::grad(pcf);

I get the following error related to the type declaration which i do not understand:
error: conversion from ‘Foam::tmp<Foam::GeometricField<Foam::Vector<doubl e>, Foam::fvPatchField, Foam::volMesh> >’ to non-scalar type ‘Foam::surfaceScalarField {aka Foam::GeometricField<double, Foam::fvsPatchField, Foam::surfaceMesh>}’ requested
surfaceScalarField newTerm = fvc::grad(pcf);
^

Though i sense it requests me to change from surface to volScalarField it makes no real meaning.

For reference i post the related code.

surfaceScalarField fc_0 = mixture.surfaceTensionForce();
fc = fvc::average(fc_0 * mesh.Sf()/mesh.magSf());

fvScalarMatrix pcEqn
(
fvm::laplacian(pc) == fvc::div(fc)
);

pcEqn.solve();

surfaceScalarField pcf = fvc::interpolate(pc);

surfaceScalarField newTerm = fvc::grad(pcf);

Appreciate any sort of help.

Thanks;
Saideep

[Saideep]

hi guys;

Just in case if anyone comes across this error!

Well i made an obvious mistake by taking the gradient of the scalar and defining it in terms of a scalar again.

Firstly "grad of a scalar gives a vector". So, it should be a vol or surface VectorField based if the value is being computed at cell center or at the face centers.

--
An another trick or rather something interesting that i came across is using a surfaceVectorField term for the computation of an surfaceScalarField.
It is indeed not possible to do so. The obvious choice again is to convert the VectorField into a scalarField and i did it using the dot product of two vectors yielding a scalar. However, this becomes quite tedious as I had to modify several terms along with their dimensions in the process.

Saideep