Steel rolling process with large deformation - implementing ALE adaptive meshing

[polak7]

I am simulating a rolling process of steel. The model is 3D with symmetry planes in Y and Z directions. I am using Abaqus Explicit for this. The height reduction of the material is high - around 80%. The simulation runs successfully, however mesh elements deform heavily, so I want to use adaptive mesh. The problem is that when I turn on ALE adaptive mesh, the simulation terminates shortly after the beginning.

First, I will show you the case without adaptive meshing.
Before simulation:


After simulation (elements are heavily distorted):


So, I turned on ALE adaptive mesh:


But the simulation terminates shortly after the start:


The results for the last successful increment are given below:


And the comparison between two simulations (with and without ALE adaptive meshing) at 0.015 step time (total time is 0.1):


It can be seen, that the quality of mesh improved.
I don’t understand, why the simulation with ALE adaptive mesh terminates. Shouldn’t it be opposite?

Some bonus images with additional information:




How can I solve this problem?

 

[FEA way]

Maybe try increasing the frequency of remeshing but also make sure that your loading is realistic - check the kinetic energy, perhaps mass scaling is too large.

 

[polak7]

The frequency is set to 1 per increment, so it can't be increased more:


The kinetic energy is relatively low:

 

[FEA way]

You could try playing with the ALE adaptive mesh controls. But the initial mesh (quality/shape of elements) also matters.

Check the documentation examples under Example Problems --> Static Stress/Displacement Analyses --> Forming analyses (especially "Flat rolling: transient and steady-state").