INVESTIGADORES
HUESPE Alfredo Edmundo
congresos y reuniones científicas
Título:
CONTINUUM MULTI-SCALE (FE2) MODELING OF MATERIAL FAILURE IN CONCRETE-LIKE MATERIALS
Autor/es:
M. CAICEDO; J. OLIVER,; A E. HUESPE; E. ROUBIN; J HERNANDEZ
Lugar:
Barcelona
Reunión:
Conferencia; XIII International Conference on Computational Plasticity (COMPLAS 2015; 2015
Institución organizadora:
CIMNE
Resumen:
Two-scale computational modeling of materials is a subject of increasing interest in computationalmechanics, particularly, two-scale computational modeling of material failure is more controversial,and exhibits additional complexity. Either if discrete approaches, or continuum approaches are used atthe lower scale, the kinematic description of some, or both, scales cannot be considered smoothanymore, and the existence of the RVE can be questioned arguing that, in this case, the material losesthe statistical homogeneity. A crucial consequence of this issue is the lack of objectivity of the resultswith respect to the size of the RVE. In (Drago and Pindera 2007) a recent attempt to overcome thisflaw, for regularized non-local models, can be found This work is an attempt to address this issue inthe setting of the Continuum Strong Discontinuity Approach (CSDA) to material failure, developed bythe authors in the past (Oliver, Huespe et al. 2002).The essentials of the proposed approach are:1. At the macroscopic level, material failure is captured via strain-localization and finite elementswith embedded regularized strong discontinuities.2. A failure-cell at the microscopic scale, with the same size and topological properties than the RVEis associated to material points at the strain-localizing part of the macrostructure. This failure-cellis enriched with appropriated material failure mechanisms with, apparently, no restriction on theirtype. Though, for the sake of simplicity, cohesive-bands with a predefined position have been usedin this work, there is no ?a priori? limitation on using more sophisticated material failuremechanisms, e.g. arbitrarily propagating cracks or strong discontinuities. In contrast, this failurecell is not claimed to be a RVE, in the sense of being statistically representative of any part of themacrostructure, although standard homogenization procedures are applied to it.The proposed approach was validated and tested with a wide number of representative examples, oneof them considered as a classical benchmark in concrete structures, this is the case of the NooruMohammed test.