[an error occurred while processing this directive]
In the fracture model
presented in this paper, the basic assumption is that the energy is the sum of two
terms, one elastic and one cohesive, depending on the elastic and inelastic part of the
deformation, respectively. Two variants are examined: a local model, and a nonlocal
model obtained by adding a gradient term to the cohesive energy. While the local
model only applies to materials which obey Drucker’s postulate and only predicts
catastrophic failure, the nonlocal model describes the softening regime and
predicts two collapse mechanisms, one for brittle fracture and one for ductile
fracture.
In its nonlocal version, the model has two main advantages over the models
existing in the literature. The first is that the basic elements of the theory
(the yield function, hardening rule, and evolution laws) are not assumed,
but are determined as necessary conditions for the existence of solutions in
incremental energy minimization. This reduces to a minimum the number
of independent assumptions required to construct the model. The second
advantage is that, with appropriate choices of the analytical shape of the cohesive
energy, it becomes possible to reproduce, with surprising accuracy, a large
variety of observed experimental responses. In all cases, the model provides a
description of the entire evolution, from the initial elastic regime to final
rupture.