[saddy]

https://www.cfd-online.com/Forums/op...o-regions.html

here our super moderator Tobias developed a fantastic extension for mesh refinement at the interface.
i am trying to compile for openfoam-v1612+
i got following error

wmake libso .
wmakeLnInclude: linking include files to ./lnInclude
Making dependency list for source file dynamicInterfaceRefineFvMesh.C
g++ -std=c++11 -m64 -DOPENFOAM_PLUS=1612 -Dlinux64 -DWM_ARCH_OPTION=64 -DWM_DP -DWM_LABEL_SIZE=32 -Wall -Wextra -Wold-style-cast -Wnon-virtual-dtor -Wno-unused-parameter -Wno-invalid-offsetof -O3  -DNoRepository -ftemplate-depth-100 -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/triSurface/lnInclude -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/meshTools/lnInclude -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/dynamicMesh/lnInclude -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/finiteVolume/lnInclude -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/dynamicFvMesh/lnInclude -IlnInclude -I. -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/OpenFOAM/lnInclude -I/home/hasan/OpenFOAM/OpenFOAM-v1612+/src/OSspecific/POSIX/lnInclude   -fPIC -c dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C -o Make/linux64GccDPInt32Opt/dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.o
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C: In member function ‘Foam::autoPtr<Foam::mapPolyMesh> Foam::dynamicInterfaceRefineFvMesh::refine(const labelList&)’:
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:349:17: error: ‘fillSignallingNan’ is not a member of ‘Foam::sigFpe’
                 sigFpe::fillSignallingNan(phi.internalField());
                 ^
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:389:13: error: ‘GeometricBoundaryField’ is not a member of ‘Foam::surfaceScalarField {aka Foam::GeometricField<double, Foam::fvsPatchField, Foam::surfaceMesh>}’
             surfaceScalarField::GeometricBoundaryField& bphi =
             ^
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:389:57: error: ‘bphi’ was not declared in this scope
             surfaceScalarField::GeometricBoundaryField& bphi =
                                                         ^
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C: In member function ‘Foam::autoPtr<Foam::mapPolyMesh> Foam::dynamicInterfaceRefineFvMesh::unrefine(const labelList&)’:
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:579:13: error: ‘GeometricBoundaryField’ is not a member of ‘Foam::surfaceScalarField {aka Foam::GeometricField<double, Foam::fvsPatchField, Foam::surfaceMesh>}’
             surfaceScalarField::GeometricBoundaryField& bphi =
             ^
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:579:57: error: ‘bphi’ was not declared in this scope
             surfaceScalarField::GeometricBoundaryField& bphi =
                                                         ^
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C: In member function ‘virtual bool Foam::dynamicInterfaceRefineFvMesh::writeObject(Foam::IOstream::streamFormat, Foam::IOstream::versionNumber, Foam::IOstream::compressionType) const’:
dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.C:1662:9: error: ‘writeObjects’ is not a member of ‘Foam::dynamicFvMesh’
         dynamicFvMesh::writeObjects(fmt, ver, cmp)
         ^
/home/hasan/OpenFOAM/OpenFOAM-v1612+/wmake/rules/General/transform:25: recipe for target 'Make/linux64GccDPInt32Opt/dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.o' failed
make: *** [Make/linux64GccDPInt32Opt/dynamicInterfaceRefineFvMesh/dynamicInterfaceRefineFvMesh.o] Error 1
hasan@hasan-Lenovo-G50-70:~/OpenFOAM/hasan-v1612+/OpenFOAM_extensions/dynamicinterfacerefinefvmesh/src/dynamicFvMesh$

i need help guys. i have to do this. i need interface refinement. i tried to look into the code but still errors are not going away.
please take a look

[Tobi]

Hi Saddy,

first of all thanks for the first sentence
To your question. Unfortunately the library is based on OpenFOAM-2.3.x and I should update it to 5.0 (maybe today in the evening). However, if you just need an interface refinement, you can use the standard library. Just set the refinement between e.g. alpha 0.4 and 0.6. In addition to your question, the field names changes during different releases and even more if different openfoam versions are used. Mine is based on the Foundation one but you are using the one from ESI. I am not sure if it is compatible. You can try to compile the library if I updated it to version 5.x but I don't think that it will work out of the box. But again, an interface refinement can be done with the normal library too.

[Tobi]

Done,

I got the same errors in the 5.x version. So now you can use the library for the Foundation Version and I guess also for the ESI one (because I had the same errors).

Cheers.

[saddy]

Originally Posted by Tobi

Done,

I got the same errors in the 5.x version. So now you can use the library for the Foundation Version and I guess also for the ESI one (because I had the same errors).

Cheers.

First of all, i said it because your work is really fantastic. kind of gerris code octree which is amazing
i have also found that openfoam lacks in refinement department. if there was anisotropic/unstructured/tetrahedral refinement available. openfoam would be very robust.
thanks so much for updating the code. it means a lot.
actually my problem is based on refining the mesh inside a cylinder and i have tried with normal library, trust me. but it doesn't work half as good as yours. i installed your extension in OF2.3.1 and the RESULTS WERE COMPARABLE TO GERRIS OCTREE.
i ran the dam break problem and i used refinement level 8 with just a single mesh consisting of two blocks only.
your library i feel makes it faster too.
i'll try installing for my openfoam-v1612+ and let you know if it works

[Tobi]

Just out of curiosity - what do you mean with Gerris Code Octree.

[saddy]

i am sure you aware of gerris flow solver. actually gerris uses octree meshing. the simulation starts with a single block and refinement takes place only where the interface is. the adaptive meshing of gerris is by far the best. your code follows the same refinement followed by gerris
there is no need for giving a base mesh like in openfoam. simple geometries can be given by means of equation: cylinder by x2+y2=r2 and extruding in z..simple
a good visulaization of gerris octree meshing is here;
https://www.youtube.com/watch?v=yJhYENolQeU
https://www.youtube.com/watch?v=3Jf_-XMv1wk

the discussion on it's tree strucure/adaptivity is given here.
in short gerris:solves for each finer subgrid to capture flow physics,to reach kolmogrov length scales
https://www.youtube.com/watch?v=DKn9iuD7Ihk

i believe openfoam lags in this department.
GERRIS adaptivemeshing comes so close to DNS!.

[Tobi]

I am sorry but I never heard about Gerris. Seems that it is not longer in development. Thanks for the information but I will not check it out. If you are so happy with that software, use it.

[alinuman15]

That is really beneficial discussion. However, I would like to add that Gerris is not that effective for handling the contact angle phenomena particularly for 3D problems. That why I switched towards OpenFoam after working with gerris for a while.
Now, I have solved very complicated cases of interface tracking. However, I never used the adaptive mesh refinement. Any idea about links, examples , ... etc preferably if there is something works with Salome?
Thank you all


Sent from my iPhone using CFD Online Forum mobile app

[akidess]

FYI: Gerris has long been superseeded by PARIS: http://www.ida.upmc.fr/~zaleski/paris/index.html

[alinuman15]

That is true ( I am following Gerris news!) but not for tracking the contact line in its 3d fashion.
Regards

[saddy]

Yes , Gerris is no longer in development.
it's been followed with basilisk which is still in process of testing and development,
paris is also a good.

but openFoam needs unstructured adaptive refinement. because adaptive meshing saves time when it comes to simulating big & complex geometric dimensions.
it's not possible to figure out snappyhexmesh problems and keep trying until you get it right.
and unless you have hexahedral mesh openfoam doesn't refine.it's just not possible to make hexahedral mesh for some geometries at all
take a look at these posts,

Adaptive mesh and movement

[alinuman15]

Thanks Saddy. I will look at take a look at that post.


Sent from my iPhone using CFD Online Forum mobile app

[saddy]

Originally Posted by Tobi

I am sorry but I never heard about Gerris. Seems that it is not longer in development. Thanks for the information but I will not check it out. If you are so happy with that software, use it.

i did use gerris earlier, but i switched to openfoam because there is no phase change in gerris. i am working on phase change.
and yeah i installed your OpenFOAM5.0 it compiled after 1 error: line1662 i changed writeobjects to writeobject. unfortunately it compiles but doesn't work on the way it should.
i'll try to sit down with the codes and try to compile it again.!!!
thanks

[saddy]

PROFESSOR Zaleski ia coming to our college in september to give guest lectures