Numerical study of self-assembly of granular and colloidal particles

Abstract

Self-assembly of granular materials and colloids are studied using several different computational methods such as Discrete Element Method (DEM), Smoothed Particle Hydrodynamics (SPH) method, finite volume Volume of Fluid and DEM (VOF-DEM) method and coupled VOF-Level Set and Dissipative Particle Dynamics (CVOFLS-DPD) method. A history dependent contact model is developed for the DEM and a cohesion model is introduced to study the packing of granular materials under cohesive forces. The study reveals granular size and size distribution has an important effect on the final packing structure. The study using SPH method reveals stress relaxation in a granular system subjected consecutive jamming cycles. However, above a certain initial packing fraction stress relaxation is found to be negligible. Further analysis reveals characteristics length and time scales for stress relaxation. Three-cycle basis is found to be the most preferred configuration of the particles as the granular system drives towards a more stable state. The study using VOF-DEM method reveals pattern formation by colloidal deposits as a thin film of fluid evaporates. Further analysis with CVOFLS-DPD method reveals interface forces on particles need to be carefully modeled to prevent escaping of particles during evaporation. The use of machine learning (ML) for computational study is also explored in this study. A machine-learned sub-grid scale (SGS) modeling technique is introduced for efficient and accurate prediction of reactants and products undergoing parallel competitive reactions in a bubble column. The machine-learned model replaces the iterative approach associated with the use of analytical profiles for previous sub-grid scale models for correcting concentration profiles in boundary layers.