I used the Tool “Thicken shell mesh” to pass from a shell mandrel to a solid mandrel in an explicit simulation.
The tool works. After redefining the section and a few other necessary moves, I tried to run it, but I got an error.
I had to try a series of actions, but at the end the problem looked like a boundary condition had been messed up by the tool, although form the interface everything seemed fine.
I deleted the boundary condition and re-generated it, and everything went fine.
This happens every time I run the “Thicken shell mesh” tool.
I cannot share the file because for some reason (independent from the thicken tool), it has grown to a huge size: 1.4 GB!!!
Do you know why the “Thicken shell mesh” tool can mess with boundary conditions?
Would be good to see how it was defined in keywords back then.
Does it include the results ? Those can be large in explicit dynamics if you don’t limit the output frequency significantly. If that’s the case, you could close the results, resave the file, and then share it here. Or use some hosting website if it’s still too big.
I will try to produce a simple input to reproduce the issue, I guess it has nothing to do with the step being explicit.
No, that’s strange for me, it is only the .pmx. The results (.frd) are 160MB.
I have 10 y experience with the explicit of Abaqus for rolling simulations, and I’m trying to understan what can be done with Calculix. While the implicit solver is really great and understandable, the explicit solver, as managed by PrePoMax remains somehow indecipherable.
Thank you FEMAnalyst!
I followed your tutorial some weeks ago. I work at PPM only when I have some spare time.
I also followed a tutorial about a simulation of a ball hitting a wall. I improved it and played with it …
What I am now trying to do is something like this, completely inside PPM:
mandrel rolling.
It works, but not always. Sometimes hourglass kicks in or funny behaviours of the solver.
The speed is very slow, if compared to Abaqus explicit, despite the large number of CPUS that I can use, and there is no clear management of the mass scaling.
It would also be helpful to have support for time-dependent plastic curves, but ok.
Now I see why you needed the Thicken mesh tool - you have rigid body constraints so explicit won’t work with shells (the GitHub issue I linked before).
Yeah, those are the quirks of explicit dynamics (sometimes even contact in general) with CalculiX.
Maybe you could consider using OpenRadioss instead.
Do you mean rate-dependent plasticity ? Recently, the JC model with rate dependence was added to CalculiX, but it’s not supported in PrePoMax yet so you would have to use the Keyword Editor to add it.
Correct. In the example above I have one 3D body (the roll) and one shell (the madrel) that I thicken.
That’s a good suggestion, thanks.
I mean simply the possibility to input different plastic curves (each with a softening and flat tail) for different strain rates. But that’s a “second” level capability, necessry only for exploring plastic instabilities due to viscous properties.
I was not able to reproduce the model. But the idea behind it is that after the mesh is thickened, all selection-based items are reselected. So this re-selection could be the reason for your problem. I would need an example to be more specific.
