Micro-mechanical analysis of quasi-static particulate fragmentation applied to geomaterials

Abstract

Despite the progress made on modeling granular mechanics at the macro-scale, discriminating the mechanisms that control particle breakage within a particulate assembly is still an open issue. In this thesis, we analyze particle fragmentation and the associated energy distribution from a micro-mechanical standpoint. First, we explain the state of the art in terms of micro-macro modeling of breakage in granular assemblies, and we present in detail the pros and cons of the techniques employed in the Discrete Element Method (DEM) to simulate particle crushing. Second, we conduct a numerical study at the particle scale to understand in which conditions shielding effects overcome size effects. We also investigate the influence of microscopic flaws on particle crushing. Our simulations highlight the predominant influence of porosity over flaw size and show that particle strength depends linearly on the particle coordination number and quadratically on particle porosity. Third, we analyze micro-CT scans obtained sequentially during an oedometer test of zeolite to model crushing at the scale of a granular assembly. We implement a new breakage model in DEM, which combines particle replacement, for primary splitting breakage, and cluster bond breakage, for modeling for fragment breakage. The PSD obtained with the new DEM model exhibited an impressive matching with experimental results. Fourth, we analyze the evolution of the energy of a granular assembly subject to particulate breakage during quasi-static confined comminution. Energy potentials are related to internal variables that have a clear relationship to microstructure evolution, e.g. deviation to initial and ultimate PSDs, particle specific surface, porosity. DEM simulations show that: At least 60% of the work input is dissipated by particle redistribution; The breakage energy accounts for less than 5% of the total input energy; The energy dissipated by redistribution is between 14 to 30 times larger than the breakage energy. Lastly, we propose a DEM displacement-softening contact model to simulate fracture propagation in concrete. We calibrate the model with results of uniaxial compression tests and Brazilian tests of both mortar and concrete. DEM simulations confirm that concrete does not fail in pure tensile mode during Brazilian tests. Sensitivity analyses also show that concrete strength depends linearly on both the adhesive area fraction of the ITZ and the aggregate tensile strength. The ITZ has a greater effect on concrete strength than aggregate strength. Future studies will focus on the modeling of various simultaneous breakage mechanisms (e.g. splitting, multi-fragmentation, chipping, erosion) and on their relation to the macroscopic response of the granular assembly, as well as the modeling of cyclic effects, the prediction of bond corrosion in the presence of water, and the design of bond reparation.