(192d) Simulating Complex Mixing Configurations within a Single Numerical Framework

Mixing flow for industrial applications often exhibits a multitude of physical phenomena such as aeration, emulsion formation, particle clustering, as well as cavern formation in yield-stress fluids. Despite the fact that enormous progress in terms of computational methods has been made for each of these phenomena, it is also rare that a single numerical framework is capable of capturing them all in detail. This talk will highlight our innovative numerical algorithms able to accurately handle: (i) complex geometrical constructions (impellers, baffled tanks, and static mixers); (ii) cavern formation when a stirred vessel agitates viscoplastic materials [Russell et al., 2019]; (iii) aeration when dealing with challenging rapid air-liquid mixing systems [Kahouadji et al., 2022]; (iv) emulsions in liquid-liquid mixing systems [Liang et al., 2022, Vald ́es et al., 2023]; and (v) particle transport in multiphase mixing systems.

Within our framework, any numerical configuration which necessitates the construction of complex geometrical shapes uses a technique that circumvents the need for time-consuming construction, meshing, and re-meshing of these geometries. Instead, our method proceeds in a modular manner that enables the user to build any geometrical shape from primitive geometrical objects (spheres, cylinders, planes, torii, and ellipsoids) using static distance functions that take into account the interaction of these objects with the flow for both single and multiphase flows.

As an example of mixing with non-Newtonian fluids, “cavern” formation occurs when a viscoplastic material is agitated by a central impeller within a vessel. Outside of this rotating region, the material is stagnant because the stresses being imparted on it are not sufficiently large to exceed its yield stress. This is important as most industrial applications need to minimise stagnant zones in order to achieve efficient, complete mixing. In this situation, our numerical framework has shown remarkable capability of reproducing experimental measurements and excellent agreement with scaling laws involving a combination of dimensionless parameters that incorporate rheological, inertial, and geometrical effects.