Localization properties of strain-softening gradient plasticity models. Part II: Theories with gradients of internal variables

The second part of this paper compares and evaluates enhancements of the conventional plasticity theory by gradients of internal variables. Attention is focused on their performance as localization limiters. Both explicit and implicit gradient formulations are considered. It is shown that certain models suffer by serious mathematical deficiencies that would complicate their numerical implementation. Some other models are appropriate only at early stages of the softening process but later exhibit locking accompanied by a spurious expansion of the localized plastic zone. The comparative study indicates that a convenient and robust tool for regularized modeling of the entire localization process is provided by the implicit gradient approach combined with a suitable form of the hardening/softening law.