Mesoscale FEM approach on cemented sands: Generating and testing the digital twin

• Published in: E3S web of conferences Vol. 544; p. 7003
• Main Authors: Komodromos, Michail, Stamati, Olga, Grabe, Jürgen
• Format: Journal Article
• Language: English
• Published: EDP Sciences 2024
• Subjects: adapted image meshing algorithm; cemented sand; mesoscale fem; x-ray ct
• ISSN: 2267-1242; 2267-1242
• DOI: 10.1051/e3sconf/202454407003

The mechanics of Cemented Granular Material (CGM) have been studied by means of geotechnical experimental testing, whose output consists the basis of mathematical models which approach the material response in various loading states. The information derived from standard experimental response curves is the basis of understanding and handling the material. Still, it is intuitive to analyse the CGM down to the mesoscale and dray conclusions over the interaction of the constituent material phases. Diverging from the practice of equivalent continuum, the alternative description of a three phase composite of sand particles, cement binder and void pores has been realised in this study. In order to implement the specific morphology of this multiphase, geomaterial, X-ray Computed Tomography is used to capture the internal structure and quantify it into a three dimensional greyvalue map (or a three dimensional image). The distinction of the material phases is made possible by the application of a developed filter, which corrects the artefacts caused by beam hardening phenomena and allows for the generation of a phase segmented equivalent image. An image adapted meshing algorithm has been utilized to transform the labelled image into a tetrahedral mesh, grouped into sets that correspond to the different materials. The tetrahedral domain was assigned boundary conditions and was numerically tested under uniaxial compression using the finite element method. The kinematics of the simulation proved that the mesoscale approach, which carries internal structure information of the granular fabric and the cement paste distribution, provides a output which captures the kinematics of the granular skeleton.