Qihan,

There are (at least) two contributions to the Bauschinger effect.

The first is grain scale differences in plastic deformation that lead to internal stress development. Upon stress reversal, those internal stresses modulate the response and can result in a non-symmetric (tension–compression) stress–strain curve. That aspect would emerge from a "plain" polycrystal simulation provided that the mechanical response of the grains in the aggregate is (sufficiently) different. In case of two-phase structures, that is very likely.

The second is what you are probably eluding to, which is the aspect of polarization of the dislocation structure, i.e. same sign dislocations concentrating in different volumes within individual grains. This causes a more localized internal stress as those same sign dislocations generally repel each other. To include this aspect, you are right in that you would have to extend the current constitutive formulations. Nevertheless, this should be a manageable effort, since there are only a few additional parameters required when extending (one of) the existing formulations. There is already one example that includes this aspect of the Bauschinger effect: phase_plastic_kinehardening is derived from phase_plastic_phenopowerlaw and can demonstrate how such additional aspects would change the sourcecode.

Dear Qihan Gao

Few years ago I made a thesis about cyclic loading and dislocation based modelling using DAMASK, it is freely available here:

https://infoscience.epfl.ch/record/223625

If you are interested in one of the models developed, I can send you the relevant code. However, it is based on an older version of DAMASK.

The backstress caused by the repulsion between same sign dislocations can be definitely modelled using DAMASK if you activate the dislocation fluxes (non-local model). Indeed, the dislocation flux can model GND accumulation and the corresponding stress that affects the neighbouring dislocations.

Best Regards,

Nicolò

Dear Nicolo,

I've read thesis about cyclic loading and dislocation based modelling based on an older version of DAMASK in https://infoscience.epfl.ch/record/223625

And I'm interested in the models developed, could you send me the relevant code? The backstress caused by the repulsion between same sign dislocations can be definitely modelled using DAMASK if you activate the dislocation fluxes (non-local model).

Best Regards,