Microstructure-informed modelling of cement mortar using XCT imaging and phase segmentation

Cement-based materials, including concrete and cement mortar, possess an inherently heterogeneous microstructure, resulting in variations in macroscopic properties such as stiffness, strength, and durability. This study introduces a microstructure-informed model for cement-based materials, which predicts mechanical behavior at Level III using constitutive models developed from data at Level II and Level I. This approach integrates microstructure information from different length scales (a multiscale framework) within a continuum mechanics framework. This approach uses micro-CT imaging and grayscale-based phase segmentation to link micromechanical properties with grayscale intensity. High-resolution 3D imaging and advanced phase segmentation techniques are employed to generate 3D digital cement mortar (DCM) for microstructure informed finite element (FE) modelling. Constitutive models, informed by nanoindentation data, are integrated into the FE model to predict compressive, tensile, and damage behaviors. The Unified Mechanics Theory (UMT) [1] is utilized to model damage evolution, employing the thermodynamic state index to assess degradation functions based on microscale entropy estimations. The results reveal that UMT effectively predicts damage progression without relying on traditional strain-based curve fitting commonly used in conventional damage evolution functions. Model predictions were validated against experimental compressive test data and showing strong agreement with Mori-Tanaka and self-consistent homogenization schemes for elastic modulus estimation. The study highlights the significant influence of microstructural heterogeneity on damage localization and mechanical resilience. The findings demonstrate that grayscale-based phase modelling, combined with UMT-based degradation model, provides a robust approach for linking micromechanical properties to global mechanical performance.