Elmer: An Open Source Finite Elements Software

29.06.2017
simulations_elmer

There are nowadays some Open Source Finite Elements packages available online, just to cite some: CalculixSalome-Meca (with Code Aster)Z88 AuroraElmer, etc.

The gap between the paid software (AnsysComsolAbaqus, LS-DYNA,…) and the open source packages is still big, especially when it comes to the GUI and the Post-Processing. It is also sometimes difficult to find documentation or useful examples which makes the learning curve quite steep.  In addition,  when working with OS software, where the code has not run through that many validation steps, it becomes essential to be aware of what is the software actually doing and always have a critical eye when analyzing the results. These reasons make Finte Elements Free packages not so popular.

However, Free FE software, have nowadays reached a quite advanced stage being perfectly capable of solving general models. Free code software becomes especially useful when a rare case is approached. Some special models cannot be implemented by commercial software since the required conditions have not been foreseen. When working with OS, it is possible to contact the current code developers and even ask them to develop special tools for your application. This is almost impossible when dealing with the big paid software packages.

Watch video!

Elmer Multiphysics

banner

Elmer is an open source multiphysical simulation software mainly developed by CSC – IT Center for Science (CSC). Elmer development was started 1995 in collaboration with Finnish Universities, research institutes and industry. After it’s open source publication in 2005, the use and development of Elmer has become international.

Elmer is actually a suite of several programs and you can use them independently.

  • ElmerGUI: Pre- and Post-processing
  • ElmerSolver: Solution
  • ElmerPost: Post-processing
  • ElmerGrid: structured meshing and mesh import. Partitioning.

pre_solver_post

Elmer counts with several Physical Models between them: Heat transfer,  Fluid mechanics, Structural mechanics, Acoustics, Species transport, Electromagnetics, Mesh movement (Lagrangian), Level set method (Eulerian), Elctrokinetics, Thermoelectricity, Quantum mechanics, Particle tracker. See Application Examples.

The main Elmer links: Official WebpageVersion Control SystemDiscussion Forum.

I personally worked with the following combination of OS giving a satisfactory work flow:

banner

Even being quite outdated, I used the Linux distribution CAELinux package and all worked like a charm. Give it a try!

Managing Editor, PhD Researcher & Founder of TriboEngineering

3 Comments

  1. “the open source packages is still big, especially when it comes to the GUI and the Post-Processing.”
    Not.
    I’m using commercial packages for FE simulation and my own converters to ParaView, because post-processor of the enterprise-grade package just sucks. They made a monster in BAT files/sh scripts/tcl/C++ and it is unstable, slow and doesn’t show all results. ParaView is open source.
    After using FE in optimization I found that pre-processing GUI is not needed if there is a documentation and examples. Optimization engine formulates input files from checked templates. The biggest problems of open-source FE tools are:
    – Lack of documentation
    – Lack of non-linear formulations for material models. In physics EVERYTHING is non-linear. Conductivity, emissivity, diffusion, reology. There are equations for it and they are used in commercial solutions since 1980s.
    – No coupled simulation
    – No incremental simulation and terrible discretization for it. Velocity-location-pressure numerical instability (solved usually by placing artificial node in the middle of the element) is intentionally ignored.

  2. Do you know if anyone has managed to implement a tribology solution, eg elastohydrodynamic lubrication in Elmer or any of these open source codes?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.