Meso-scale Fracture Modelling Of Concrete Based On Unified Phase-field Damage Model

Concrete is a kind of heterogeneous and quasi-brittle composite material,and its failure is the most common engineering problem.In this paper,heterogeneous mesos-copic models are established.Three approaches combining unified phase field model(PFD)are developed to model complex microcracking in heterogeneous cementitious materials at meso-scale.The proposed numerical model allowed us to predict accurately in crack nucleation and propagation in concrete,in order to investigate the relationship between the meso-phase and the macroscopic behaviour,and to accurately understand the complex damage and fracture mechanisms of concrete materials.In a first part,the numerical simulation tool,unified phase-field damage model(PFD)is developed and introduced.More specifically,PFD is expressed in a unified form based on the thermodynamic law.The characteristic functions can be optimized to the regularize cohesion zoom model(CZM).PFD can be used to simulate the fracture behavior of brittle and quasi-brittle materials.Then,in order to study the feasibility of the unified phase field damage model to simulate the fracture damage of heterogeneous materials at the meso-scal,new numerical simulation methods for the mesoscopic models are developed.Firstly,the method combining X-ray computed tomography(XCT)images and PFD is considered.By means of XCT situ experiments and image processing,the real meso-scale structure of concrete specimens can be obtained.The self-programming MATLAB program is used for pixel-based meshing,and finally the finite element model of the real meso-structure can be obtained.The simulations of the meso-structure are conducted with PFD,and the sensitivity of the method to mesh size and length scale in the heterogeneous model are verified.Secondly,meso-csale models of Weibull random field are established to describe the random non-uniformity of material strength.Then PFD is applied to simulate the damage of highly heterogeneous meso-scale structure.The impact of parameters in the random field are analyzed.Finally,the preliminary study on the reconstruction of random media based on XCT images is conduct.Based on the real image of XCT,the geometric information of aggregates can be extracted by the statistical equivalent representative volume element(SERVE).Based on this information,a series of relevant random aggregates can be established.Moreover,based on nonlinear transformation of Gaussian field,equivalent non-Gauss random fields can be established according to the marginal distribution and auto-correlation function of XCT samples.Monte Carlo simulation is used to verify the stability and feasibility of these models.The thesis carries out theoretical derivation and numerical simulation,aiming at opening up a new method of mesoscopic numerical model characterization,proposing new ideas for meso-scale concrete simulation,and providing reference for the study of concrete fracture mechanics at multi-scale.