Except for reading messages such as warnings and errors, it’s better to check the Status and Convergence tabs first (in this order since the Status tab provides quick insight into incrementation progress).
The model may still be underconstrained. You could try adding soft springs to the blank but then they should be removed with *MODEL CHANGE in the later stages of the simulation. Ot you could try dynamic implicit step with some significant damping but it may not converge either.
Generally speaking, CalculiX is not really a good choice for simulations like this. As I said before, they are typically carried out with explicit dynamics so OpenRadioss would make sense. Even if the initial phase converges in CalculiX, you will likely encounter issues when the deformation becomes large.
Hello @DrHybris,
after many adjustments, I was able to run the simulation almost to completion (it finally stopped after 36mm). However, I reduced the blank holder force to 4.9kN, since you applied the full force of 19.6kN to the 1/4 model.
Unfortunately, reducing the blank holder force in my simulation caused the blank to lift the holder and form wrinkles (i think for this reason i needed to lower the contact stiffness further from 10xE to 1xE to make the solver converge):
@Gunnar , I have tried without success running your linked ModeloSolid_mod.pmx with the default PrePoMax “PrePoMax v2.3.1 dev”
so I would please like to ask which version of Calculix you have been used for your shown example ?
Has it been the default following with “PrePoMax v2.3.1 dev”
or
has it been the default from (https://www.dhondt.de/calculix_2.22_4win.zip)
or
have you downloaded and compiled the currrent version from (GitHub - Dhondtguido/CalculiX: This repository contains the source files of CalculiX, a three-dimensional Finite Element Program (www.calculix.de).)
or maybe you have your own modified version ?
My reason for asking is that with the default version of PrePoMax v2.3.1 dev, solver and your dataset the Calculix solver exits after only 3 steps with too many cutbacks, so I will expect something must be different between your setup and a default setup from PrePomax
SUMMARY OF JOB INFORMATION
STEP INC ATT ITRS TOT TIME STEP TIME INC TIME
1 1 1 6 0.100000E+01 0.100000E+01 0.100000E+01
2 1 1 44 0.200000E+01 0.100000E+01 0.100000E+01
3 1 1U 8 0.200000E+01 0.000000E+00 0.600000E-03
3 1 2U 60 0.200000E+01 0.000000E+00 0.600000E-03
3 1 3U 60 0.200000E+01 0.000000E+00 0.300000E-03
3 1 4U 60 0.200000E+01 0.000000E+00 0.150000E-03
3 1 5U 60 0.200000E+01 0.000000E+00 0.750000E-04
3 1 6U 38 0.200000E+01 0.000000E+00 0.375000E-04
The analysis has only 3 steps. But yes, it fails to converge at the beginning of the third step. For me, it looks like this with PrePoMax v2.3.0 and ccx 2.22 provided with it:
Btw. It might be worth mentioning there’s a limitation in CalculiX where rigid body + shell + Nlgeom results in non-convergence. So it may indeed be better to use extruded single layer solids (shells are extruded to them anyway).
I have linked ccx outside prepomax so i am not using the ccx version which is part of the prepomax package (and i didn’t tested it yet).
But have just downloaded the version from here again (https://www.dhondt.de/calculix_2.22_4win.zip). After adding the required DDL files for the Pardiso solver, I ran the simulation briefly without problems.
Hm… I just noticed that it works on my PC at work (Intel Xeon W-2225 / Windows 11 Pro ), but NOT on my old PC (Intel i7-6700 / Windows 10 Home). I use always 4 cores and testet it with ppx v2.3.0 and v2.3.1, it works with both versions with there includet ccx versions. So it seems to depend either on the windows version or the hardware… Are you using Windows 10 or 11?
Unfortunately, I’m no longer so optimistic. If I increase the blank holder force to avoid warping, this phenomenon occurs at some point
Also if I replace the blank holder force with a fixation (first animation), it causes the sheet to jam.
I don’t think it makes a big difference, but I’ll try to simulate this later without friction.
