Mr Hui Li

Complex crack propagation at micro/meso-scale in heterogeneous early-age concrete is usually induced by non-uniform shrinkage and thermal expansion during hydration processes, directly affecting the loading-carrying capacity of concrete structures and their systems. Prediction of such early-age fracture is essential for investigating its effects on the macroscopic mechanical performance of concrete and further optimizing structural design. To this end, this study proposes a novel mesoscale hydration-induced fracture modelling method combining a chemo-thermo-mechanical coupled phase-field model and random aggregate models for complex mesoscale early-age concrete fracture simulations. In this method, the Fourier's law and the Arrhenius's law are used to simulate heat transfer and hydration reaction in heterogeneous models, respectively. The temperature and hydration degree of solids are fully incorporated into the governing equations of the phase-field regularized cohesive zone model to efficiently simulate complicated chemo-thermally induced fracture, without the need of remeshing, crack tracking or auxiliary fields. The resultant displacement-temperature-hydration degree-damage four-field coupled system of equations is solved using a staggered Newton¿Raphson iterative algorithm within the finite element framework. The new method is first verified by a heat convection problem with numerical solutions and a hydration fracture problem of a concrete ring with experimental data. Mesoscale fracture modelling of an early-age concrete square is then carried out to investigate the effects of mesh size, phase-field length scale, boundary conditions, and the distribution and volume fraction of random aggregates, on concrete hydration. It is found that the present method is capable of accurately and robustly modelling chemo-thermally induced mesoscale multi-crack propagation, with insensitivity to mesh size and phase-field length scale. The capacity of modelling complex heterogeneous early-age cracking, as well as its potential for advancing structural design and optimization, is well demonstrated.

This study develops two mesoscale hydraulic fracturing modelling approaches, combining a hydromechanical coupled cohesive phase-field model with random field (RF) models and random aggregate models, respectively, for simulating complicated 2D/3D mesoscale hydraulic damage and fracture in quasi-brittle materials. The hydromechanical cohesive phase-field model is used to automatically model fluid-driven crack initiation and propagation in the generated mesoscale models without remeshing. Fracking experiments of a concrete cube under hydraulic pressure were simulated and analysed as benchmark examples. Monte Carlo simulations with 2,400 RF-based samples were first carried out to investigate the effects of variances, correlation lengths, and confining stresses on fluid flow and crack propagation. It was found that a higher variance of tensile strength led to lower mean peak pressures, while a larger correlation length and confining stress had an opposite effect. Random models consisting of mortar, polygonal/polyhedral aggregates, and interfacial transitional zones (ITZ), were then simulated, and the predicted linear peak pressure-confining stress relation was found in good agreement with the experimental and published numerical results. Although more validation is needed, the capacity of simulating complex and realistic fracking processes and the potential for engineering fracking design and parametric optimization of the developed approaches is well demonstrated.