Font Size: 

In this study, continuous random fields of material properties (e.g. tensile strength) are generated based on real microscale X-ray Computed Tomography (CT) images of concrete, through statistical reconstruction algorithms and the Karhunen-Loève expansion technique. The random fields are then mapped to finite element models, before mesoscale crack initiation and propagation are simulated by a phase-field cohesive zone model (PF-CZM) with softening laws. The proposed modelling method is validated by Monte Carlo simulations of a square specimen under uniaxial tension and an L-shaped panel under mixed-mode fracture. It is found that the method can capture realistic mesoscale fracture processes and predict reliable macroscale load-displacement responses. The developed method provides potential for the efficient evaluation of structural reliability considering real microstructures.

mesoscale concrete; CT image-based modelling; random field; phase field model; damage and fracture