I have checked this web site many times , but I can not find
The package described as âCalculiX with all solvers and DLLs from the IntelÂŽ oneAPI Math Kernel Library 2022.3 required by the Pardiso solver.â should have it.
1 Like
Thank you for support
Any additional action required to switch to ooc version ?
Best regards
I know this is not a solution but, looking at your mesh, the mesh density is much higher than needed for an engineering application. I would try to decrease the mesh element size first. Even if you can run such a model, the pre and post-processing will be much more problematic/slow. You can increase the mesh density on local regions using Mesh refinements.
Dear Matej
I wrote earlier that this structure consist of steel plates , looks like giant box , thatâs why I can not use low density mesh .
Best regards
You can always start with a coarser mesh and perform mesh convergence study - refine the mesh and check how the results change. At some point the results should stop changing significantly and further refinement wonât increase the accuracy anymore.
when I use low density mesh I obtain negative Jacobian
1 Like
Negative Jacobian usually occurs when the mid-nodes of the parabolic tetrahedral elements are projected to the geometry. Are you using projection of the mid-side nodes on geometry = Yes.
1 Like
where can I check it or switch off ?
thank you for help
Does it affect results quality ? , since I have read that parabolic elements should have 10 nodes ?
It doesnât decrease the number of nodes, itâs just about projecting the midside ones to the geometry so that its curvature is better captured.
thanks it helps a lot but the message in PPM mesher window is a little bit ambiguous 
Hello Michael!
Here is a good description. It helped me understand the meshing parameters.
http://www.tu.kielce.pl/~rokach/instr/PPM_instalacja.htm