Calibration of spherical harmonic-based random fields from μCT images to generate shot-earth 3-D aggregates

Shot-earth is an eco-friendly material that has been recently rediscovered in many applications within the building sector. However, its mechanical characterization presents many challenges, due to its complex internal structure, which is characterized by the presence of aggregates with highly variable shapes. This study aims to develop and validate a computational framework for the probabilistic modelling of shot-earth aggregate geometry. Realistic 3-D aggregates are generated using spherical harmonic-based random fields combined with Monte Carlo simulations. While these stochastic and mathematical tools have previously been applied in the meso-scale modelling of concrete-like materials, their application to shot-earth has not been explored yet. The generation process is informed by dimensional data obtained from micro-computed tomography (μCT) images of real cubic shot-earth samples. Statistical analysis of these data is used to set the assumptions for the construction of a spherical harmonic-based random field and infer the probability distributions of the required coefficients. The resulting virtual aggregates are compared to the original ones in terms of volume, moments of inertia, and overall shape. Quantitative comparisons indicate good agreement with the experimental data, demonstrating the effectiveness of the proposed framework. This approach enables more accurate meso-scale simulations of shot-earth aggregates and supports future developments in finite element modelling for studying their mechanical behaviour, ultimately promoting the broader adoption of shot-earth in sustainable construction. Moreover, the detailed workflow provides a reproducible methodology that can be generalized to other granular materials, from μCT data to the generation of virtual samples.