Elmer Discussion Forum

Dear Elmer Users,

We are happy to announce a new Elmer minor release 8.2.

The source code of the release is available in "release" branch of GitHub at https://github.com/ElmerCSC/elmerfem. You may retrieve the release at any later date using its tag by "git checkout release-8.2". There is also a fresh Windows installer for version 8.2 available at

https://sourceforge.net/projects/elmerfem/.

The release notes list the most important new features of Elmer since the last official version 8.1 published on 29 October 2015. There has been around 300 code commits after that time.

For most parts the release notes cover only features that are visible for the end-user, or a developer utilizing only some default routines. Beyond that there has been numerous improvements and fixes to the code that are not directly accessible concern to the end-user. Therefore we have not covered them in the notes.

Still the most active application fields have been computational electromagnetics and computational glaciology (Elmer/Ice). For the latter community the information of recent developments is made available via Elmer/ICE user community. Note that there are some large active branches that will be merged to devel after this release to gradually make it to the next official release.

Elmer documentation has also been modified to better follow the features of version 8.2. Unfortunately the documentation is never quite complete and is still lacking behind in some developments.

We would like to thank the developers outside CSC for many important contributions. The contributors for this release based on the git log information have
been: Eelis T., Joe T., Juha R., Juhani K., Laure T., Mika M., Mikko B., Pavel P., Peter R., Rupert G., Sami I., and Thomas Z.

If your name is not on the list even though you have contributed your contribution may be waiting in some of the branches.
Also there has been many users testing Elmer and sending bug reports. Without these efforts development of Elmer could not have been as fast.

The release notes are attached here for convenience.

Best regards,
ElmerTeam
CSC - IT Center for Science, Finland
http://www.csc.fi/elmer

Hello dear developers,

quite excited to read the release notes of version 8.2.

Specially the point:
o Allow combination of 2D and 3D heat solver by adding a keyword for the 2D part thickness

got my attention.

Does this mean one can have a mesh now where we combine volume element bodies (e.g. tets) with face element bodies,not bcs (eg.triangles)?
And can this also be a combination?
As in a body of triangles that also forms the bc of a volume?
This could for example be used to solve a heat equation in 2d on the boundary of a box.

best regards
Franz

Hi

Yes. Even previously it has been possible to have other equations solved with given boundaries, but then boundaries only. The problem is more specific when you are combining 2D and 3D. This particular extension was for HeatSolver when you have, for example, a conducting layer on top of a 3D insulating body but don't want to mesh for the thin layer. Then adding it an effective thickness may do the trick.

-Peter

Hi,

is this feature available for other solvers also? I would be particularly interested in having it for Static Current Solver, for simulation of a thin metallic layer on top of other material.

Matthias

HI there,

so how would i realize a mesh with mixed elements?
I just tried a naive approach where i meshed a rectangle and a box in salome and i exportedi t to unv.
then with elmergrid i converted to elmer format.

but here i only have 504 elements (tets) left in the mesh.bodies file.

best regards
Franz

Hello,
me again on the topic of 2d and 3d mixed meshes. Is there an minimal example somewhere on how to realize such a mesh?

or cna anybody summarize a how to here?

best regards
Franz

Hi

It depends. Would you like to have the lower dimensional mesh to be 1) a thin layer on top of the higher dimensional one or 2) lower dimensional continuation of the mesh in ~normal direction.

For heat equation these test cases could provide some limited guidance (fresh devel branch) for the above cases
1) ThermalPipe3D
2) MortarMixedDimensions

-Peter

Hi Peter,

is there any chance to extend this functionality to electric, e.g. statcurrent?

Thanks,

Matthias

NIce,
thank you peter!
i will have a look at them. But first i need a coffee.
Its my how-to-whatever when it comes to computerwork: coffee.

best regrads
Franz

Hi all,
I am going to try and revive this old post. I have recently opened a thread asking about 2D/3D meshes in static electric current solver (viewtopic.php?f=7&t=5976), are there any plans to add this as it was done in the heat solver?

Thanks!