On the various numerical methods for the simulation and validation of thermovibrationally-driven solid particle accumulation phenomena in microgravity conditions

New microgravity-based contactless solid particle manipulation strategies relying on the application of thermal stimuli to a fluid in combination with imposed high-frequency vibrations offer promising directions for the production in space of new materials that cannot be obtained in normal gravity conditions. This line of inquiry, originating from a theory formulated ten years ago, has recently received confirmation by virtue of experiments conducted on board the ISS using the Selectable Optical Diagnostic Instrument (SODI) in conjunction with the Microgravity Science Glovebox (MSG). These experiments have also revealed interesting effects, which were not included in the original theory and need now to be properly interpreted, understood and modeled. In such a context, this study is devoted to a critical comparison of different available Eulerian-Lagrangian numerical strategies potentially able to reproduce the experimental findings and reveal in detail the underlying mechanisms. These differ in regard to the degree of fidelity with which the interactions between the suspended solid particles and the carrier fluid are implemented. Different numerical approaches with increasing complexity are used as the discussion progresses. Starting from the simplest possible case where the back influence of particles on fluid flow is disregarded (one-way coupling), we then consider the standard two-way coupling model by which the liquid and solid phases are fully coupled in terms of momentum exchange and the Dense Discrete Particle Modelling (DDPM) where in addition to the localized exchange of momentum, the inter-particles stresses are also somehow taken into account by analogy with the kinetic theory of gases. Finally, a four-way coupling strategy is implemented, where effective particle collisions are also simulated (through proper coupling of Ansys Fluent and Rocky software). We show that different approaches display a varying degree of success in reproducing properly the experiments, which depends on the ability of the solver to capture properly particle inter-stresses.