Around five years ago I wrote some simple things for Calculix. Like a searchable documentation and a web based way for solving small models online. You can browse those and some other tools here:
Due to lack of time and interest I never updated them… However, I was thinking to update everything now and make some enhancements. The enhancements would be:
- Ability to submit small and medium size models to run for free on my own cluster ( with some limitations )
- Ability to visualize results with full 3D controls using the FRD to VTK converter and ParaView for the web…
- Write an installation script for the Pastix solver
Any interest in any of this?
What is medium size? Also FRD to VTK and Paraview with the cluster might be good. Install script for Pastix might also be good. This all depends on the definition of “medium”. I currently use Mecway and will continue to do so but might find paraview helpful, And I have peers at work with no easy access to FEM.
Hi, @feacluster that is a great initiative! The installation script for Pastix solver would be so helpful to many newcomers.
The use of the FRD to VTK converter is a very good idea
I have used your repository for the CCX Verification Examples, very useful when don’t wish to download (or restricted).
Thank you. Look forward to seeing new features.
I have completed my updates and you can check them out here:
Main feature is the ability to submit a deck to run for upto 30 minutes using google cloud compute optimized machines . Other things are:
My focus now will be to make these existing tools more robust. Once that is done I will look into offering the pastix solver and the frd to vtk translator ( to post process smaller models through the web ).
I’ve implemented your frd->paraview translator to my web-based calculix solver here:
After solving a small model , click the button to visualize it in 3D. It should bring it up in ParaView Glance like this:
Thank you! I think you’re brilliant. And I think there is too much honor for myself - URL to the project on GitHub would be enough.
I tried to post process frd within Paraview, but it seems Paraview cannot distinguish parts without inp model information. Therefore it treats the frd as a single and compound part. Is it correct?