Particulate flows are prevalent in nature and industrial processes, showcasing complex dynamics. This thesis presents a coupled fluid-particle simulation method to investigate these systems at the microscale by fully resolving relevant physical effects. It integrates the lattice Boltzmann method with the discrete element method to capture fluid flow dynamics and model frictional collisions among rigid particles. The study emphasizes the characteristics of the coupling approach, comparing various techniques to enhance simulation results. An extensive calibration and validation pipeline ensures accuracy, focusing on explicitly modeling short-range hydrodynamic interactions and adequately resolving single collision events. This verification concept is applicable to fully resolved simulation approaches, providing guidelines for model parameterization. All components are optimized for efficient massively parallel execution on supercomputers, crucial for managing the high computational demands of microscale studies. Performance improvements stem from innovative dynamic load balancing techniques that address the varying workloads. The method is employed to simulate the erosion and transport of sediment beds composed of tens of thousands of particles, yielding detailed data that enhances the understanding of macroscale system dynamics. This work contributes significantly to the multi-scale modeling framework of particulate flows.
