Material softening and convergence issues 3

[hermeneutics]

Hey everyone,

I am modeling an isotropic material with an elastic-plastic behavior. I am applying a prescribed displacement. I want to soften the material but the *STATIC analysis never converges. I tried *Static, Riks but it never reaches the prescribed displacement. As soon as the material reaches the yield point, the displacement starts to decrease and I never get to see the softening. I've tried "automatic stabilization" with the *Static procedure but I couldn't get it to run past yield. I also tried breaking it into several steps with a small displacement increment but it still failed. The error is always the same, "time increment required is less than the minimum specified" even though my time increment is currently 1E-10.

Any advice is appreciated,
Thanks!

 

[mrgoldthorpe]

Try relaxing the displacement correction criterion using *CONTROLS ....

Once the load/displacement increments are very small the "noise" in nodal displacement corrections during iterations of an increment can easily be the same size as the nodal displacement increments themselves. This leads to further cut-backs and the error you experience.

Furthermore, set an initially small time increment in the load step, and don't allow the maximum increment in a step to become too large.

The following sets a more relaxed displacement tolerance of 10%, i.e. allows corrections to the displacements of 10% compared with the actual displacements.
*CONTROLS, PARAMETERS=FIELD, FIELD=DISPLACEMENT
** R_n^alpha C_n^alpha q_0^alpha q_n^aplha R_p^alpha eps^alpha
, 0.10 , , ,

Restricting the maximum size of the increment:
*STATIC
** init'l time period min. max.
** time of step time time
0.01 , 1.00 , 1.0E-6 , 0.01

 

[mrgoldthorpe]

... by the way: the reason for restricting the maximum size of an increment.

If you let ABAQUS increase it by default, and then it's forced to cut-back because it's rather too large when (say) gross yield or softening cuts in, you can end up completely losing the advantage gained from the previous larger increments.

The slow and steady tortoise often wins against the stumbling hare ....

 

[hermeneutics]

Thanks for your help. It helped a bit, but not to the degree of softening that I was looking for. I implemented the Ductile Criterion damage initiation to soften it and found that was a bit more robust than my "manual" softening. Still, had to settle with a "slow" damage.