Liuchi Li, Applied Mechanics, Ph.D. Thesis Defense
"Linking micro-structure to macro-behavior of granular matter: from flowing heterogeneously to morphing adaptively"
From concrete gravels unloaded from trucks to wheat seeds discharged through funnels, from polymeric beads filled in shoe cushions to metallic pellets packed in robotic grippers, granular matter is becoming increasingly relevant in coping with our evolvingly sophisticated societal needs in many respects (e.g. expanding urbanization, growing population and advancing manufacturing). This increasing relevance urges developing micro-structural understandings of granular matter regarding their two basic macro-scale behaviors: flowing heterogeneously and morphing adaptively. However, findings in this regard so far suffer from a disconnection in length-scale - some adopting a top-down perspective lack predictability due to few insights taken from underpinning micro-scale details (e.g. microstructures), while others adopting a bottom-up perspective lack practicality due to few specificities incorporated from overlaying macro-scale conditions (e.g. heterogeneities).
In this dissertation, via Discrete Element Method (DEM) simulations, we bridge the divide between length-scales in this regard by revealing the fundamental role of microstructures. To begin with, we evaluate and verify the robustness of DEM in capturing granular microstructures, by systematically comparing simulation results with experimental measurements on quasi-statically sheared granular assemblies. Then, we first numerically study spatial phase transitions in heterogeneous granular flows from a top-down perspective. We start by calibrating and validating a DEM model using experiments we perform on fluidizing spherical particle pile formed in a rotating drum, we next take the validated model to produce flows with different microstructures by systematically varying boundary condition and loading rate, and lastly we study their correlations with phase transitions ranging from gas-like layers near the free surface, to underneath liquid-like layers, and to solid-like layers deep in the bulk. We propose a micro-scale parameter quantifying the level of structural anisotropy, that can for the first time elucidate the spatial phase transitions between these layers in dependent of imposed boundary conditions and loading rates. Further, we find that, in solid-like layers, this micro-structural quantity correlates to bulk effective friction, an integral macro-scale quantity in constitutive modeling. Next, we numerically study bending modulus adaptations in shape-morphing granular sheets from a bottom-up perspective. We start by calibrating and validating a DEM model using experiments we perform on bending 3D printed granular sheets enclosed in a flexible membrane, we next take the validated model to construct granular sheets with different microstructures by varying constituent particle shape, initial configuration and confining pressure. Lastly we study the correlation between microstructure variations and modulus adaptations. We discover a universal power-law correlation between bending modulus (a macro-scale quantity) and coordination number (a micro-scale quantity) in reminiscence of the canonical power-law scaling for packing of frictionless sphere near jamming. We also find larger coordination number favors interlocked particles over non-interlocked ones, leading to significantly better shape-morphing performance of chain-like sheets over discrete assemblies.