The flow of granular materials is important in many manufacturing processes, especially in the pharmaceutical, energy, and agricultural industries, and in naturally-occurring processes such as avalanches, landslides and sediment transport. Modeling the behavior of flowing granular materials is challenging because observable measurable properties of the flowing material depend on microscopic dynamics on the scale of the individual grains of the material. In the absence of reliable models, the design of equipment to handle granular materials is often based on heuristic approaches, which are not optimal and occasionally lead to catastrophic failures. This award will support a new framework for predicting the dynamics of granular systems that addresses small-scale and large-scale phenomena at the same time. The new framework will be a hybrid approach that combines models of granular materials as fluids with models of granular materials as collections of discrete particles. The key challenge that will be addressed is to discover the appropriate way to combine these approaches in a hybrid model that accurately simulates well-defined prototype granular flows and is validated against comparisons with industrial and geological scale granular flow phenomena. Computer codes that are generated in the project will be shared publicly, and the research team will engage students at various academic levels to participate in the project, especially students from traditionally underrepresented groups.
A hybrid framework for simulating granular flows will be developed to obtain the efficiency of continuum simulations at the macroscale and the accuracy and precision of discrete particle simulations offer at the microscale. Four components of the hybridization scheme will be explored: a coupling procedure that mechanically handshakes the continuum regime and the particle regime along a reconciliation zone, a homogenization operator that replaces a discrete element zone with a continuum state, an enrichment operator that acts in the opposite direction, and an oracle that determines where it is safe to homogenize and where enrichment to discrete grains is needed. Components of the scheme will be separately tested and validated against analytical results and experimental data. The overall hybrid model will then be validated against large-scale granular phenomena with the goal of accurately simulation flows up to the industrial- and geo-scales. The results of this work and its associated computer codes will be made available to practitioners and researchers in industry and geosciences and to developers of cinematic realizations of granular flow phenomena.
