[dariodario32]

Hi all,

I'm facing a problem with the mesh that I'm trying to generate. My purpose is to generate a conformed mesh in the fluid-structure interface. Thus, I want to know if there is a way to generate the fluid mesh with snappyHexMesh, specifying the faces on the boundary (extracted from the structural mesh). In this way the structural and fluid mesh will match in their interface.

THanks

[ashish.svm]

Originally Posted by dariodario32

Hi all,

I'm facing a problem with the mesh that I'm trying to generate. My purpose is to generate a conformed mesh in the fluid-structure interface. Thus, I want to know if there is a way to generate the fluid mesh with snappyHexMesh, specifying the faces on the boundary (extracted from the structural mesh). In this way the structural and fluid mesh will match in their interface.

THanks

Were you able to create the mesh from the extracted surfaces? Please let me know, I am trying to do the similar thing.
Thanks.

[dariodario32]

As far as I know there is now way to do it in SHM. Thus I highly suggest you to move to other mesher, have a look to SALOME (open-source).
It allows you to create meshes with different elements, ranging from 1D to 3D.

Dario

[ashish.svm]

Originally Posted by dariodario32

As far as I know there is now way to do it in SHM. Thus I highly suggest you to move to other mesher, have a look to SALOME (open-source).
It allows you to create meshes with different elements, ranging from 1D to 3D.

Dario

Can you please explain briefly, how you created conforming mesh for a given surface in SALOME? Or can you refer a similar online tutorial case by which I can learn to create similar mesh

[dariodario32]

Originally Posted by ashish.svm

Can you please explain briefly, how you created conforming mesh for a given surface in SALOME? Or can you refer a similar online tutorial case by which I can learn to create similar mesh

My solution was the following:

1. Create the geometry with a CAD software and import it in SALOME or draw it in SALOME. This geometry should contain both fluid and structure geometrical domainS. In case your fluid domain will be discretized with 3D elements, while the structure domain with 2D elements, i.e. when your FSI domain is a mix of volumes and surfaces, you use the compound object in SALOME. It is convenient to add the geometries related to the fluid domain in a group called FLUID and the ones belonging to the solid doamin in a group called SOLID.
   Then you generate the mesh of the compound geometry using the SALOME internal mesher. Automatically the mesh at the interface between the solid and structure domains will be a conformed mesh.
2. Then you can generate groups of elements starting from the groups generated in the geometry. In this way you can export the two sub-meshes in different files.
3. The fluid mesh can be now converted in OpenFOAM using ideasToFoam command.
4. The solid mesh can be imported in the an OpenFOAM solver in a similar way. You can have a look to the code of ideasToFoam application (or similar applications for mesh conversion) and integrate it into your code.
5. Lastly I guess you will need to map the stresses and/or displacements at the interface. For this task I coded a pre-processing application that compare the fluid and structure mesh in a way that you know which node correspond to which and which face to which (this is required since when the two meshes are exported separately in SALOME, the numbering of nodes and elements at the interface is not the same, hope this explaination is clear).

In case you have any question, please ask.

[ashish.svm]

Originally Posted by dariodario32

My solution was the following:

1. Create the geometry with a CAD software and import it in SALOME or draw it in SALOME. This geometry should contain both fluid and structure geometrical domainS. In case your fluid domain will be discretized with 3D elements, while the structure domain with 2D elements, i.e. when your FSI domain is a mix of volumes and surfaces, you use the compound object in SALOME. It is convenient to add the geometries related to the fluid domain in a group called FLUID and the ones belonging to the solid doamin in a group called SOLID.
   Then you generate the mesh of the compound geometry using the SALOME internal mesher. Automatically the mesh at the interface between the solid and structure domains will be a conformed mesh.
2. Then you can generate groups of elements starting from the groups generated in the geometry. In this way you can export the two sub-meshes in different files.
3. The fluid mesh can be now converted in OpenFOAM using ideasToFoam command.
4. The solid mesh can be imported in the an OpenFOAM solver in a similar way. You can have a look to the code of ideasToFoam application (or similar applications for mesh conversion) and integrate it into your code.
5. Lastly I guess you will need to map the stresses and/or displacements at the interface. For this task I coded a pre-processing application that compare the fluid and structure mesh in a way that you know which node correspond to which and which face to which (this is required since when the two meshes are exported separately in SALOME, the numbering of nodes and elements at the interface is not the same, hope this explaination is clear).

In case you have any question, please ask.

Hi Dario, Thanks for your reply. Now I can create conforming mesh using salome but there is one problem. Number of cells generated are very large. As you can see the following:
veryFineMesh.jpg
here black lines represent the fluid region. This is actually generated in another code and exported as stl in salome. Following is the stl file
fluidSTL.JPG
The above image may look like surface but it is actually a solid with very small width.

In salome meshing I am using "partition" method for geometry. And for meshing I am using automatic 3D tetrahedralization.

Is there any way I can generate coarse mesh for the region away from the fluid region?

[dariodario32]

You can try to solve this generating sub-meshes in SALOME. To do this you'll need to generate groups on your geometry, for example you can group the edges of your domain. Then right click on your mesh and generate a sub-mesh on the groups generated. In the case of the edges you can assign 1-D hypotesis (this allows you to distribute a number of elements with constant or non-constant distribution, here you decide the number of elements).

You can even create sub-meshes on surfaces.

With this method SALOME will first generate the sub-meshes and then the internal mesh in a way that it fulfill the sub-mesh hypotesis.

[ashish.svm]

Originally Posted by dariodario32

You can try to solve this generating sub-meshes in SALOME. To do this you'll need to generate groups on your geometry, for example you can group the edges of your domain. Then right click on your mesh and generate a sub-mesh on the groups generated. In the case of the edges you can assign 1-D hypotesis (this allows you to distribute a number of elements with constant or non-constant distribution, here you decide the number of elements).

You can even create sub-meshes on surfaces.

With this method SALOME will first generate the sub-meshes and then the internal mesh in a way that it fulfill the sub-mesh hypotesis.

Thanks Dario.
I tried to create submesh for the boundary faces and defined large grid size there. But it helped in reducing only very few number of cells.
Actually it will be better if I can generate mostly hexahedral mesh. Is there any way to generate mixed mesh (consisting of hex,tetrahedral, prisms etc) in SALOME?

[dariodario32]

Yes you can generate mixed mesh. As before you can group your domain in sub-domains and assign to each sub domain a different meshing solver, hexa , tetra etc.

[HakikiCanakkaleli]

Hi,

I might misunderstand (possibly did not understand) your objective. Yet the lumped point boundary condition might be of help to your purpose?

Lumped point Fluid Structure Interaction (FSI)