Résumé
Silo analogies are commonly used in block caving research to study gravity-driven flow and stress redistribution. However, most existing models rely on idealized granular assemblies composed of random granular packings, which might differ from fragmented rock masses. The aim of this numerical study is to investigate the influence of material characteristics on flow dynamics and internal stresses in a silo. Using the Discrete Element Method (DEM), we compare three silo models: (i) a caving-like (CL) configuration, generated by progressively de-bonding a tessellated solid column during draw, thus emulating rock fragmentation; (ii) a caving-polygons (CP) configuration, in which the same fragments are deposited randomly into the silo; and (iii) a caving-disks (CD) configuration, representing a traditional granular silo composed of circular particles; all three models share the same particle size distribution. These configurations enable comparing the effects of granular packing, particle shape, interlocking, and fragmentation mechanisms on discharge behavior. Results show that flow velocities, displacement fields, and internal pressure distributions differ markedly between traditional granular silos and fragmentation-driven systems. The CD configuration develops the highest pressures surrounding the outlet, whereas the CL configuration redistributes stresses more efficiently, with forces preferentially supported by the silo walls; the CP system exhibits intermediate behavior.The findings highlight key limitations of conventional silo analogs and underscore the need for more representative material descriptions in block caving simulations. This work provides a basis for selecting appropriate DEM modeling strategies, with perspectives of improving DEM cave-flow predictions.