ElmerGrid not able to convert from any format to .grd format.

[granzer]

I looked at the actual mesh files and the problem is that boundary 3 is associated with both the solenoid and the surrounding air
Boundary 2 is associated with both the coil and the surrounding air. In Elmer a boundary can be associated with two bodies.
I manually edited the boundary file to associate b3 with only the solenoid and b2 with only the coil, and the GUI behaves as you expect it.
I suppose when a boundary is tied to two bodies the GUI does not know which one to select.

For example boundary element 47011
47011 3 6171 0 303 1258 1243 1260
label, boundary, parent1 parent2 type n1 n2 n3
I changed parent2 to 0 it was assigned to an element in body 3
parent 1 is in body 1
select1.png
select2.png

Does setting the Parent 2 to 0 break the connections between Parent 1 and Parent 2...that is will the magnetic field generated by the coil still be able to permiate the air domain? I am asking because not the coil boundary is not having any connection to parent 2 by making it 0?

[raback]

Hi

1) Internal boundary may have two parents
2) When reading in the boundaries their parent information is verified
3) I guess in this case the internal BC is not even used to anything
4) ElmerGUI is more picky than ElmerSolver. I don't really have much experience on internal BCs on it.

-Peter

[kevinarden]

"Does setting the Parent 2 to 0 break the connections between Parent 1 and Parent 2...that is will the magnetic field generated by the coil still be able to permiate the air domain? I am asking because not the coil boundary is not having any connection to parent 2 by making it 0?"

Elmer shares data across internal boundaries through shared nodes, no boundary is needed. If the meshes are independent and do not share nodes, than a boundary for each body is needed in order to use mortar boundary conditions to share data.

[granzer]

Hi

1) Internal boundary may have two parents
2) When reading in the boundaries their parent information is verified
3) I guess in this case the internal BC is not even used to anything
4) ElmerGUI is more picky than ElmerSolver. I don't really have much experience on internal BCs on it.

-Peter

Ok. In my model the boundaries of Plunger and Coil should act as internal boundries right and so they should have 2 parents. How is it the they are not used...as you are stating in 3?. Also, if I set parent 2 to 0 will it work and still maintain the mesh connection between coil/Plunger (internal boundries) and the air domain??

[kevinarden]

The boundary between the coil and the air is made by the shared nodes at the interface.

[granzer]

"Does setting the Parent 2 to 0 break the connections between Parent 1 and Parent 2...that is will the magnetic field generated by the coil still be able to permiate the air domain? I am asking because not the coil boundary is not having any connection to parent 2 by making it 0?"

Elmer shares data across internal boundaries through shared nodes, no boundary is needed. If the meshes are independent and do not share nodes, than a boundary for each body is needed in order to use mortar boundary conditions to share data.

Does that mean if I delete the surface mesh and only keep the 3 volume meshes ..before exporting from Hyperemesh it will work?

[granzer]

No deleting the surface mesh and keeping only the volume mesh dosnt work.....i.e mesh will load but have the same problem of not being able to select the bodies individually

[raback]

Hi

Internal boundaries are not needed to ensure continuity. They could be used to apply some additional flux but my imagination fails me to suggest what that could be in this case. Still, sometimes they can be useful for postprocessing. In this case maybe not so much because the postprocessed elemental fields are discontinuous and the boundary values are hence not well defined. For these kind of applications I often use "Save Bulk Only = Logical True" for the vtu output because the boundaries do not really carry any useful additional info.

-Peter

[granzer]

Hi

Internal boundaries are not needed to ensure continuity. They could be used to apply some additional flux but my imagination fails me to suggest what that could be in this case. Still, sometimes they can be useful for postprocessing. In this case maybe not so much because the postprocessed elemental fields are discontinuous and the boundary values are hence not well defined. For these kind of applications I often use "Save Bulk Only = Logical True" for the vtu output because the boundaries do not really carry any useful additional info.

-Peter

I am sorry if this question seems redundant, I am asking just to make sure I understand correctly. So if I edit the boundary file and set the parent 2 to 0, as you mentioned, it wont create any problem in sharing data at the internal boundary region and I will be able to simulate by setting the equations to the correct bodies and get the results?

[raback]

Hi

Yes, the internal boundary without a BC is not used for anything in ElmerSolver.

ElmerGUI seems to use it to highlight objects. That's certainly a deficiency.

What is your original mesh format? Do you have "mesh.names" in your mesh directory?

-Peter