Speaker
Description
At a fundamental level, the macroscopic response of granular media depends on the spatial organization of contact forces between grains—the so-called ‘force chains.’ Despite their critical importance, force chains in granular media have been characterized and analyzed almost exclusively in 2D systems. To address this knowledge gap, we recently developed a new approach: a tomographic imaging technique (interference optical projection tomography, or IOPT), which by combining the principles of photoelasticity and tomography, provides direct visualization of the particles’ force network, thus circumventing the need of constitutive models of particle-particle contact (Li and Juanes, 2024). With our novel experimental technique, we provide the microscopic explanation for why a pack of angular particles is stronger than one of round particles: they form interconnected force networks that are less likely to buckle when under stress than the isolated chains in a pack of round particles.
While early results show the potential of our approach, currently this new technique is limited to reconstructing the 3D scalar field of stress-anisotropy under axisymmetric stress conditions, for example, triaxial shear. Here, we present the reconstruction of the grain-scale full tensor field in 3D (stress-tensor tomography) and focus on the study of the 3D internal stresses in a single particle subject to arbitrary loading conditions. We use IOPT and numerical simulation to study the grain-scale frictional and frictionless contacts among particles of various shapes, such as spheres, cylinders, asperities, and half-space, and a wide range of stiffness. The forward model is used to develop a large learning set to train a neural-network representation of the tensor field. The solution to the inverse problem is enabled by incorporating the physics of the problem (balance laws and constitutive laws; e.g., Haghighat et al., 2021) in the framework of operator learning. If time permits, we will present early results extending the experimentation, modeling, and inversion of the stress field from the single-particle scale to the ensemble-scale of 3D granular packs with up to ~100 particles.
The ability to interrogate the grain-scale stresses in granular media will enable new understanding of granular media and help predict the behavior of fluid-coupled granular media in landslides, liquefaction, and earthquakes.
| References | Li, W., & Juanes, R. (2024). Dynamic imaging of force chains in 3D granular media. Proceedings of the National Academy of Sciences, 121(14), e2319160121. Haghighat, E., Raissi, M., Moure, A., Gomez, H., & Juanes, R. (2021). A physics-informed deep learning framework for inversion and surrogate modeling in solid mechanics. Computer Methods in Applied Mechanics and Engineering, 379, 113741. |
|---|---|
| Country | USA |
| Acceptance of the Terms & Conditions | Click here to agree |
