Mesoscale analysis of concrete earth mixtures, from CT to random generation of RVE

CT analysis of concrete-like materials is commonly used to generate highly accurate geometric models, an essential step in studying the characteristics of these heterogeneous materials. However, the intricate nature of the reconstructed shapes complicates the direct parameterization of their geometry, which is essential for generating numerical simulations of their stochastic aggregate distribution. This paper presents the results of ongoing research, including CT scans and numerical reconstruction of concrete-earth samples. The size and shape characteristics of inclusions identified in experiments (such as volume and inertia tensor) have been analyzed using statistical methods. Using the probabilistic characterization of the inclusions, a procedure to generate random shot-earth representative volume elements (RVE), consisting of a matrix and ellipsoidal inclusions, was developed. The samples are subsequently modeled using finite elements for Monte Carlo simulations to obtain, through homogenization, the mechanical parameters of the material and potentially reveal correlations between the elastic properties and the geometric characteristics of this innovative eco-friendly material.