Search

Continuum models of fracture should be equipped with a localization limiter to prevent pathological discretization sensitivity of finite element simulations. Moreover, stiffness recovery, called also crack closing effect, should be reproduced in the modelling of quasi-brittle materials when reversed loading is applied. In this paper the efficiency of two different established damage-plasticity models for concrete is assessed from the viewpoint of the two above-mentioned aspects. The first description is the so-called concrete damaged plasticity model, available in the ABAQUS package. In this model the plasticity theory is not only augmented with stiffness degradation and recovery, but also with the crack band approach and viscoplastic regularization. The gradient-enhanced damage description, which is the second one considered, can also be coupled to plasticity. In this model additional averaging equation prevents the pathological discretization sensitivity of fracture simulations. Two fields, displacements and averaged strain measure, are interpolated and suitable finite elements are programmed by the authors in the FEAP package. Basic concepts of the numerical analysis of the crack closing phenomenon are reviewed. Both models are capable of avoiding mesh-sensitivity, but it is achieved in a different manner and it is shown in the paper that the users of ABAQUS should employ the crack band options with particular care. Representative examples in the context of the examined issues are demonstrated: uniaxial tension of a bar and a cantilever beam. The results obtained using both the models are compared and the effects of regularization and crack closing are illustrated. [ABSTRACT FROM AUTHOR]

This paper introduces a numerical simulation tool using the Finite Element Method (FEM) for near-surface mounted (NSM) strengthening technique using fibre reinforced polymers (FRP) applied to concrete elements. In order to properly simulate the structural behaviour of NSM FRP systems there are three materials (concrete, FRP and the adhesive that binds them) and two interfaces (FRP/adhesive and adhesive/concrete) that shall be considered. This work presents the major details of a discontinuous-based constitutive model which aims at simulating NSM FRP interfaces implemented in the FEMIX FEM software. This constitutive model was adapted from one available in the literature, originally employed for fracture simulation in meso-scale analyses of quasi-brittle materials, which is based on the classical Flow Theory of Plasticity combined with fracture mechanics concepts. The most important features of the implemented constitutive model are the consideration of both fracture modes I and II and the possibility of performing 2D and 3D analysis. In the end, results based on FEM simulations are presented with the aim of investigating the soundness and accuracy of the constitutive model to simulate NSM FRP systems’ interfaces. [ABSTRACT FROM AUTHOR]

This paper compares the stress and strain state simulated in a twin-disc test and a full scale wheel/rail experiment. The simulation is done by two-dimensional finite element models. The simulations regard the scaling factor between these experiments. This factor includes the contact patch size and the deformation depth. Same stresses and strains are calculated for both experiments considering this scale factor. For the investigation of cracks in these experiments, initiated short cracks or manually produced initial cracks are needed. The transferability of the crack growth predictions from twin-disc tests to the full scale wheel/rail experiments affords a scaling of the initial crack size to the size of the contact patches. Out of practical reasons usually the initial crack tips are located in both experiments in another depth as the scaling would afford and therefore, the tips are located in another zone of the plastic deformation. Very short cracks in the twin-disc tests are removed by wear or are influenced by the always existing surface roughness. It is shown that stresses and strains scale with the experiments but this will be not the case for wear and surface roughness. [ABSTRACT FROM AUTHOR]