Concrete is a primary construction material for infrastructure; however, the progression of internal damage is difficult to estimate due to its nonlinear mechanical behavior, especially under compression stress. FEM-based simulation has been investigated as an approach to this issue; simulation accuracy is significantly influenced by the material phase distribution and material properties assigned to the model. In this study, X-ray computed tomography (CT)-based modeling is applied to simulate the compressive fracture behavior of concrete cores. The effects of material phase distribution and material properties on simulation accuracy are evaluated.

Testing concrete core samples were collected from an in-service irrigation structure in Japan. X-ray CT scanning was performed for concrete cores. The CT images are segmented into five material phases: background, cracks, voids, aggregates and mortar. Finite element meshes are generated from the segmentation data with the interfacial transition zone (ITZ) between aggregates and mortar defined as a separate phase. Compressive fracture simulations are conducted using software, with appropriate material models assigned to each phase. In addition, the influence of the presence of cracks, ITZ, and voids on simulation accuracy is examined by comparing models with and without these phases.

Thus, the deteriorated pier exhibits the highest crack volume fraction, while the less-damaged pier shows the lowest void fraction, indicating higher material density. These trends are consistent with observations on the physical core samples, demonstrating that X-ray CT-based mesh generation has the potential to reproduce material phase distributions with high fidelity. Furthermore, it is indicated that the presence of cracks and ITZ, as well as the assigned material properties, substantially affect simulation accuracy, highlighting the importance of appropriately defining material phase distribution and material properties in FEM-based fracture simulations.