Keynote Nin Yang

Mesoscale Modeling of Multiphase Flows in Stirred Tanks and Rotor-Stator Devices

Multiphase flows in stirred tanks and rotor-stator devices are characterized by complex interactions across multiple scales, from molecular-level interfacial phenomena to device-scale turbulence. Accurate prediction of key performance parameters, such as gas holdup, flow regime transition, and droplet size distribution, remains a significant challenge for conventional CFD models. This talk presents a mesoscale modeling framework that bridges the gap between sub-grid interfacial physics and macroscopic transport phenomena.

We first evaluate the role of drag models in gas-liquid stirred tanks, demonstrating that the dual-bubble-size (DBS) drag models, rooted in the Energy-Minimization Multi-Scale (EMMS) concept, significantly improve predictions of flow regimes and gas dispersion compared to classical or turbulence-corrected drag models. The formation of secondary liquid circulation near the free surface is shown to be closely linked to non-uniform gas holdup, underscoring the importance of mesoscale constraints. Extending to liquid-liquid emulsification in rotor-stator devices, we integrate the EMMS approach into CFD-population balance modeling (PBM) to derive a breakage rate corrector without empirical fitting. This corrector, which increases with rotational speed and dispersed phase fraction, enables accurate prediction of droplet size distribution, Sauter mean diameter, and span for surfactant-free systems.

Finally, we address the second mesoscale-emulsifier adsorption at droplet interfaces-through coarse-grained molecular dynamics (CGMD). By coupling CGMD-derived adsorption parameters (maximum adsorption density, diffusion coefficient, kinetic constants) with surfactant transport equations in CFD-PBM, we develop a unified model that captures the inhibition of coalescence due to interfacial coverage. This cross-scale coupling, from molecular adsorption to device-scale turbulence, represents a paradigm shift toward predictive emulsification modeling. Overall, this talk highlights the necessity and significance of crossing two distinct mesoscales-interfacial and device-level-to achieve a complete, parameter-free description of gas-liquid and liquid-liquid flows in mixing systems.