Duplicate surfaces between bodies in multiple assemblies

[karlheinzElmer]

Hello to everybody,

as a dentist I am interested to investigate the stress distribution within teeth under functional loads (linear elastistics for the beginning).

I generate the model data, which are like most medical 3D data freeform data, with a microct. The grey values are segmented (ImageJ/Fiji or ITK-SNAP) and the label masks serve as entry data for the 3d meshing using the "iso2mesh toolkit" (Matlab/Octave based). The data preparation is surprisingly easy to preform when compared to all the major players software suites (Simpleware, Mimics etc.).

As I could not find a proper export/import routine which works for "multiple bodies" (a tooth consists of enamel = the outer hard shell, dentin = the inner core and the pulp = the soft tissue inside) I decided to write an export filter which generates the Elmer mesh-format. The nice thing is: I can easily generate the mesh for Elmer for the different bodies in one run without much user interaction. The code is posted in the iso2mesh mailing list:

https://groups.google.com/forum/?hl=en& ... AmZ0vy9Zl0

However, currently I have a problem to understand a detail concerning the mesh format of Elmer.

In this post

viewtopic.php?f=7&t=2627#p8022

I found this information from Matthias Zenker: "Beware of duplicate surfaces between bodies". This, however, is excately how I exported the boundaries at the moment as I do not know the implication of duplicate surfaces for Elmer.

It would be an easy way to filter the data and remove duplicate surfaces with my export filter.

So my questions are:

1) What is the consequence of duplicate surfaces between two bodies (they share the same nodes)?
2) Do I need a boundary at all between two bodies for the solvers or the post processing?

Just to remind you - I am a dentist by training - and I started to learn FEA with Elmer just now.

As far as I understand the logic of bodies and boundaries, the bodies get the material properties, and the boundaries get the boundary conditions. If I do not need to interact which the boundary between the tissues "enamel" and "dentin" do I really need a boundary between them at all? Or would it be just enough to have the elements next to each other.

3) Are the boundaries/surfaces necessary for the display of the data ? When I use surfaces the model looks good in ElmerGUI. When I just use the elements, the data set display is not so nice. It this just "GUI cosmetics" or is there a consequence for FEA behind that?

I tried to find an answer in the documentation, but I might have missed it as well.

By the way: I decided to use Elmer as I was impressed by Elmer GUI and how easy it is to handle. Thanks a lot for sharing this great tool with us.

Sincerely

Karl-Heinz

[mzenker]

Hi,

to my understanding, all surfaces limiting bodies (in 3D) are termed boundaries, even if you do not define boundary conditions on all of them.
1) The consequence of duplicated boundaries between bodies are that they do not share nodes, so no heat or electrical current can flow between the bodies. As I do electrothermal simulations, that is a problem for me. That said, I don't know if this is also a problem in a purely mechanical simulation.
2) 3) I cannot tell either if a FEM simulation is possible with the volume elements only if you don't need to define boundary conditions.

HTH,

Matthias

[karlheinzElmer]

Dear Matthias,

your information is very helpful. Thank you for the quick reply. I will add an option to reduce the duplicate surfaces.

Sincerely

Karl-Heinz

[karlheinzElmer]

Hello everybody,

I wanted to update my recent information. As I was not sure how important it is for Elmer that the boundary between two connected bodies consists of just one surface I modified my export program for iso2mesh so that the user can select now how the mesh files should be written. By default all duplicate surfaces between bodies are unified to one surface now.

You can find the program here: https://groups.google.com/forum/?hl=en& ... AmZ0vy9Zl0

Sincerely

Karl-Heinz