Session 8 - Theme: Mixing & Processes (2)

Mixing time prediction in stirred tank reactors (STRs) remains computationally demanding when using conventional CFD approaches, limiting their practical application for design optimization and sensitivity studies. This work presents the first experimental validation of recurrence CFD (rCFD) methodology for mixing analysis in a 30L STR equipped with a dual-impeller configuration (pitched blade turbine and Rushton turbine). The rCFD approach exploits the pseudo-periodic nature of impeller-driven flows to extrapolate passive tracer transport using a pre-computed flow field database, bypassing repeated solution of the Navier-Stokes equations during the extrapolation phase. Validation was performed by comparing rCFD predictions against lattice Boltzmann method (M-STAR), finite volume method (Fluent), and experimental conductivity probe measurements at a fixed injection location. The 95% mixing time criterion was evaluated at two probe volumes and across the overall domain. Computational benchmarking revealed that rCFD requires approximately 6 seconds of clock time per second of simulated process time, compared to 240 seconds for M-STAR and 10,200 seconds for Fluent representing a speedup of nearly four orders of magnitude over conventional FVM. This computational efficiency enabled a parametric study encompassing 90 injection configurations (10 angular positions at 36° spacing, 3 radial distances, and 3 heights), a scope infeasible with traditional CFD. Results identify optimal injection locations for minimizing mixing time and reveal the sensitivity of mixing performance to injection position. The findings demonstrate rCFD as a viable tool for rapid screening and optimization of mixing processes in STR applications.

NETmix is a network mixer composed of interconnected chambers and channels that promote high heat and mass transfer rates, making it suitable for a wide range of applications, including precipitation processes. Recently, NETmix has emerged as a promising technology to produce magnesium ammonium phosphate (struvite) crystals, a sustainable alternative to conventional synthetic fertilisers. This process requires mixing three liquid streams: one containing phosphorus and nitrogen, typically originating from wastewater treatment plants; one consisting of a concentrated magnesium (Mg) solution; and one consisting of a pH-controlling solution. Two concentrated streams (Mg and pH-controlling solutions) are separately injected at low flowrate ratios relative to the main feed flowrate directly into selected chambers along the NETmix reactor. This work assesses the effect of injection chamber position on the mixing performance of the reactor. A NETmix geometry with 17 rows was used and passive tracer mixing experiments were simulated with water and two tracer species as working fluids. The topology of injection was found to have great impact on mixing. Micromixing performance was quantified using the intensity of segregation. For all configurations studied, the intensity of segregation decreased along the NETmix network, indicating progressive homogenisation of the streams. The optimised injection scheme reduced the intensity of segregation by 99.9 % and achieved comparable mixing degrees for both tracers at the reactor outlet. These results are highly relevant for crystalliser design, since mixing and in particular micromixing directly influence supersaturation, chemical reaction, and other molecular-scale processes that occur during precipitation processes.

The mechanism of turbulent mixing in liquid-phase neutralization reactions in a continuous-feed semi-batch stirred-tank reactor, where the base solution was fed with a jet in an acid solution, was investigated. The overall rate constant of the neutralization reaction, k_aV, was experimentally quantified using pH visualization images, a pseudo-first-order reaction model, and a material-balance equation for the base. The effects of the impeller rotational speed N, base feed flow rate Q, base inflow nozzle diameter din, and nozzle location on k_aV were investigated. Mathematical relationships between k_aV and Q, din, and N were derived based on micro- (engulfment) and meso- (shedding) mixing models. Two neutralization experiments were conducted: (i) continuous base injection without stirring (jet mixing only) and (ii) continuous base injection with stirring (jet and stirring mixing). In jet mixing only, k_aV was determined by the turbulent dissipation rate ε, indicating that the reaction rate is governed by the turbulent eddies elongating and entraining the surrounding unreacted fluid owing to viscous deformation (micro-mixing). For jet and stirring mixing with the nozzle set at the impeller tip, k_aV was determined by stirring-induced micro- or meso-mixing, depending on the strength of ε and the turbulent energy k in the discharge flow from the impeller. When the nozzle was positioned far from the impeller tip, k_aV was affected by both jet and stirring mixing. This effect depends on the injection and circulating flows in the reaction zone, which vary with the nozzle position.

Porous nanosilica particles are widely used in various applications due to their high surface area and tunable surface properties. Bio-inspired silica (BIS) synthesis offers a sustainable route for producing such materials under ambient conditions. In this process, a solution of sodium silicate is mixed with acidic medium (liquid, like HCl, or gaseous, like CO2). This causes precipitation, an intrinsically fast reaction, rendering the process sensitive to transport processes like micromixing and mass transfer. Previous studies have demonstrated influence of bulk pH, precursor concentration, and reactor configuration on critical properties of BIS. In this work, we investigate influence of micromixing (for liquid acidic streams) and mass transfer (with gaseous CO₂) on pH control, yield, and productivity for continuous synthesis of BIS particles. The synthesis was carried out using three fluidic devices – pinch tube, fluidic oscillator, and vortex-based cavitation device (VD), covering a wide range of micromixing times. We observed that surface area and particle size are strongly dependent on micromixing-controlled acidification rates for the HCl-based synthesis route. In addition, the effect of bubble size for the CO2-mediated silica synthesis is investigated by introducing VD for generation of fine bubbles. Enhanced gas-liquid mass transfer and reduced bubble size lead to smoother pH evolution and measurable changes in product properties, highlighting the importance of interfacial area as a design variable. Overall, this work establishes micromixing and mass transfer as tunable levers for controlling BIS characteristics and provides rational framework for designing scalable, continuous synthesis processes.

Accurate prediction of particle size distribution (PSD) is critical for the design and scale-up of crystallization processes, often encountered in fine chemical manufacturing. This remains challenging due to the complex coupling between crystallization kinetics and local hydrodynamics. Conventional crystallization models often assume perfect mixing, neglecting spatial heterogeneities in turbulence and particle–flow interactions. These simplifications limit the reliable prediction of process dynamics and PSD, particularly during scale-up and in systems where mixing-driven phenomena such as secondary nucleation are dominant. Although coupled computational fluid dynamics–population balance equation (CFD–PBE) models can capture hydrodynamic effects, their practical use is often hindered by high computational cost. This study investigates the effect of mixing on batch crystallization using a computationally efficient CFD–PBE framework based on compartmental modeling. Turbulent flow in agitated batch crystallizers is described using Reynolds-Averaged Navier–Stokes (RANS) simulations in OpenFOAM, providing spatially resolved fields of velocity, temperature, and turbulent energy dissipation. Millions of CFD cells are coarse grained to 10s of compartments, assuming perfect mixing within each compartment, using a physics-driven clustering strategy based on local energy dissipation rate. PBEs are solved over these compartments using an in-house finite volume solver. In addition, Euler–Euler multiphase CFD simulations are used to assess fluid flow characteristics under highly suspended conditions commonly encountered in industrial crystallization. Case studies show that hydrodynamic heterogeneities can significantly influence crystallization kinetics and PSD evolution. The proposed framework captures these effects at minimal additional computational cost, enabling routine use for crystallization process design and scale-up.

The optimal operation of mixer-settler systems requires the conflicting requirements of the two unit operations to be balanced. This project aims to address this challenge by establishing a closed-loop dynamic process control based on a data-driven model and inline and online drop size distribution (DSD) and conductivity measurements. The control methodology was demonstrated in a batch STR using a toluene/NaOH system, with benzoyl chloride as the transfer component to study the influence of mass transfer from the dispersed to the continuous phase. Stirrer speed was the only dynamic control variable employed to influence DSD. The dynamic behaviour of DSD in response to variations of stirrer speed was characterised in terms of coalescence and breakage time constants. An online particle detector was integrated based on an endoscopic, inline optical measurement technique enhanced with automated image analysis via a convolutional neural network. This enabled the real-time acquisition of a DSD every 2–7 s based on 4,000 detected drops. A higher-order dynamic model was identified and validated to describe the key DSD characteristics. In combination with the implementation of the particle detector and a standard PI controller, the stirrer speed could be successfully adapted in a closed-loop system. The STR setup is now operated continuously and has been extended with a settler to validate the potential of the developed methods to improve the overall robustness and efficiency of the complete mixer-settler system. This talk will provide an overview of the experimental results and modelling approaches.