[DanPad07]

Hi Foamers,

I've got an interesting problem. I want to implement a non-uniform heat flux to the surface of my model to heat up the fluid. To do this, I used a codeStream, with an excerpt of the changeDictionary dict shown below:

T
{
    internalField   uniform 323;

    boundaryField
    {
        ".*"
        {
            type            zeroGradient;
            value           $internalField;
        }
        "procBoundary.*"
        {
            type            processor;
        }
        "block_to_.*"
        {
            type            compressible::turbulentTemperatureCoupledBaffleMixed;
            Tnbr            T;
            kappaMethod     solidThermo;
            value           $internalField;
        }
		"maxZ"
		{
			type            externalWallHeatFluxTemperature;
			mode            flux;
			q               nonuniform #codeStream
			{
				codeInclude	#{
								#include "fvCFD.H"
							#};
				codeOptions	#{
								-I$(LIB_SRC)/finiteVolume/lnInclude	\
								-I$(LIB_SRC)/meshTools/lnInclude
							#};
				codeLibs	#{
								-lmeshTools \
								-lfiniteVolume
							#};
				code		#{
								const IOdictionary& d = static_cast<const IOdictionary&>
								(
									dict.parent().parent()
								); // access current directory
								
								const fvMesh& mesh = refCast<const fvMesh>(d.db()); // access mesh database
								const label id = mesh.boundary().findPatchID("maxZ"); // label id of patch
								const fvPatch& boundaryPatch = mesh.boundary()[id]; // access boundary mesh information using label id
								const vectorField& Cf = boundaryPatch.Cf(); // access coordinates of cells
								scalarField q(boundaryPatch.size(), 0); // initialise q field
								
								const scalar c0 = 0.015; // definition of the centre x coordinate
								const scalar c1 = 0.045; // definition of the centre y coordinate
								const scalar rpipe = 0.015;
								const scalar Qeff = 1e6; // Effective incident heat flux (W/m2)
								
								forAll(Cf, faceI) // loop over all the patch faces
								{
									scalar x = Cf[faceI].x(); // x coordinate of the faces i
									scalar y = Cf[faceI].y(); // y coordinate of the faces i
									scalar radius = pow(pow(x-c0,2)+pow(y-c1,2), 0.5); // compute radius on face i.
									
									q[faceI] = scalar ( Qeff*(0.5+pow(radius/rpipe,2)) ); // define heat flux value value on the face i. This is a test equation.
								}
								
								os << q;
							#};
			};
			thicknessLayers ();
			kappaLayers     ();
			kappaMethod     solidThermo;
			value           $internalField;
		}
    }
}

When solving the model in series it works fine. The issue is if I want to use multiple processors, where I get the following error when decomposing the mesh prior to using the solver:

Number of processor faces = 81286
Max number of cells = 24355 (21.72641087% above average 20007.98333)
Max number of processor patches = 6 (48.76033058% above average 4.033333333)
Max number of faces between processors = 3552 (31.09268509% above average 2709.533333)

Time = 0
Using #codeStream at line 53 in file "/Specimen_1p00/parallel/0/block/T.boundaryField.maxZ"
Using #codeStream with "/Specimen_1p00/parallel/dynamicCode/platforms/linux64Gcc63DPInt32Opt/lib/libcodeStream_7d92f667a36b3b9d505ee89b569a984a51b7824e.so"


--> FOAM FATAL IO ERROR: 

    token following 'nonuniform' is not a compound
    on patch maxZ of field T in file "/Specimen_1p00/parallel/0/block/T"

file: /Specimen_1p00/parallel/0/block/T.boundaryField.maxZ

    From function Foam::genericFvPatchField<Type>::genericFvPatchField(const Foam::fvPatch&, const Foam::DimensionedField<Type, Foam::volMesh>&, const Foam::dictionary&) [with Type = double]
    in file genericFvPatchField/genericFvPatchField.C at line 117.

FOAM exiting

I've tried removing nonuniform, but then the heat flux is not applied to the model. I've tried adding List<scalar> and other variations but to no avail. Does anyone have any tips or can catch out an error? I'm currently using OpenFOAM v1812.

Thanks in advance!

[yueyun]

Hi Dan,

I am using OF10 and I just encountered the same problem. I am afraid this reply is too late but I hope it could be helpful for others having the same problem.

I think this is because the write function of a scalarList has a different behavior. For example, if we print a scalarList, we get "n (a0 a1 ... an-1)". While if we print a vectorList, we get "List<vector> n (v0 v1 .. vn-1)".

So, please try to replace the "os << " nonuniform " << q;" with "os << " nonuniform List<scalar> " << q;" in your codeStream.

Best wishes,
Yueyun