Hey friends,
I am trying to apply ductile damage in a bending cyclic test on a pin. Using Static General Step.
I can run the simulation till the end without damage, but when I apply the ductile damage to the material, I get converge problems. It only runs until the elimination of the first elements and then gets aborted, with the usual message of “Too many attempts made for this increment”.
This is a simulation of an experimental test, I wish to simulate the full failure of the pin.
Looking at the job diagnosis, some contact occurrences: Overclosures, Maximum contact force error, Maximum penetration, points now slipping, points now sticking warnings show for Nodes around the eliminated elements, and warnings like zero pivots, and negative eigenvalues are shown.
I apply the ductile damage parameters, Fracture Strain, and Damage evolution in tabular form (D,displacement) that I simplified to just two lines (D=0 and D=1).
On the step module I am using NGeom: On and automatic stabilization, with dissipated energy fraction of 0.01.
I am using Reduced Integration linear elements C3D8R, with Stiffness Hourglass Control and max degradation of 0.998.
I think the problem is either with the damage parameter definition or the mesh definition.
Can somebody help me? I have read and tried of everything, but I can’t seem to get pass this problem. If you have a question regarding the model definition I will be glad to answer.
Thank you very much, Diogo Cabrita.
Dear Diogo Cabrita,
Thank you for sharing this discussion. This is a common error and I have encountered this problem many times. One of the reasons for this error is the lack of enough space on the computer hard drive, in addition to the incorrect definition of properties may also cause this error. I suggest making sure there is enough free space, check the properties, and finally use a supercomputer. In addition, this error may be fixed by changing the size and type of mesh. I suggest you reduce the mesh size.
Best Wishes!
How about zero stiffness when damage increases for mechanical equilibrium ?
Dear Diogo,
You do not show the timestep that you are using.
Try reducing the timestep and let us know what is the minimum timestep reached by the solver and what is the maximum strain that you aim to achieve and what is the expected strain increase per timestep.
If this quantity is too large, the simulation may not converge.
Best Regards,
Nicolò Grilli
Thank you all, I fixed the problem by using fracture energy instead of displacement in damage evolution. Also I let it run even if the abaqus froze and it goes to the end. I have another problem now that is the sitfness degradation on the unloading phases, it seems to unload elastically when the elements are already damaged and I wanted it to unload with a less stiff rate. I don't know how to consider this, if anyone has a suggestion I would like to know. Thank you again, Diogo
- Badges
- Science topic
- Similar topics
- Papers