Lattice Boltzmann study of mass transfer for two-dimensional Bretherton/Taylor bubble train flow

Alexander Kuzmin, Michał Januszewski, Dmitry Eskin, Farshid Mostowfi, J.j. DerksenChemical 2013 · DOI

TL;DR

This paper investigates mass transfer in a two-dimensional bubble train flow using Lattice Boltzmann simulations. The authors develop and validate numerical procedures to accurately determine the volumetric mass transfer coefficient, addressing challenges in defining characteristic concentrations and boundary conditions, particularly for higher capillary numbers where flow patterns become more complex.

What Problem Does This Paper Address?

The Bretherton-Taylor regime, characterized by a train of equally sized bubbles flowing through a liquid, is crucial for applications like monolith reactors due to enhanced mass transfer. However, accurately quantifying this mass transfer, especially as flow conditions change (e.g., capillary number increases), is challenging. Existing analytical models often rely on simplified bubble shapes and assumptions about mixing, which break down at higher capillary numbers where hydrodynamic patterns, like the presence or absence of a vortex in the liquid slug, significantly alter mass transfer. This paper aims to establish robust numerical procedures using Lattice Boltzmann methods to accurately determine the volumetric mass transfer coefficient across a range of capillary numbers, bridging the gap between experimental observations and simplified theoretical models.

Methods

The authors employ the Lattice Boltzmann Method (LBM) for their simulations, specifically using a two-dimensional, nine-velocity (D2Q9) TRT (two-relaxation-times) collision operator for improved accuracy in diffusion and convection. They first perform multiphase simulations to capture bubble shapes and hydrodynamic patterns for various capillary numbers (0.1 < Ca < 1.0). Subsequently, mass transfer simulations are conducted using different boundary conditions (periodic, open) and varying numbers of unit cells. A key aspect of their approach is the development of a velocity improvement technique using a single-phase hydrodynamic simulation with a free-surface solver to obtain a velocity field consistent with the advection-diffusion equation, which helps to avoid numerical instabilities and speed up simulations. They also explore velocity scaling at constant Peclet number to reduce computational demand.

Datasets

Not discussed in the paper. The study relies on numerical simulations rather than experimental datasets for training or evaluation.

Key Insights and Ideas

  • Characteristic Concentration is Key: The choice of characteristic concentration (e.g., domain-averaged vs. inlet/outlet-averaged) significantly impacts the calculated mass transfer coefficient, with domain-averaged concentration proving more robust and consistent, especially for complex flow regimes.
  • Hydrodynamics Dictate Mass Transfer: The presence or absence of a vortex in the liquid slug, which is dependent on the capillary number, fundamentally alters the mass transfer behavior and the validity of simplified analytical approaches.
  • Numerical Procedures Matter: Establishing clear numerical procedures, including appropriate boundary conditions and the number of unit cells simulated, is crucial for accurately mimicking experimental observations and obtaining reliable mass transfer coefficients.
  • Velocity Scaling for Efficiency: Velocity scaling at a constant Peclet number is a viable strategy to significantly reduce simulation time, provided the diffusion coefficient is adjusted accordingly, without compromising the accuracy of the results.

Key Results

  • The study establishes a procedure for determining the volumetric mass transfer coefficient for 2D Bretherton/Taylor bubble train flow for capillary numbers between 0.1 and 1.0.
  • They demonstrate that time-dependent volume-averaged concentration provides the most robust results for the characteristic concentration.
  • The paper shows good agreement between their simulation results and analytical predictions for different Peclet numbers.
  • Velocity scaling can achieve speedups of up to 10-40 times while maintaining simulation stability and accuracy.
  • Simulations with multiple unit cells (6-8 cells) are shown to be necessary to avoid boundary influences and provide results consistent with experimental observations.

Limitations and Future Work

  • The authors acknowledge that their work focuses on two-dimensional simulations, and future work could extend this to three dimensions.
  • While they compare their results with analytical correlations, direct comparison with experimental measurements for bubble train flow between parallel plates is limited due to a lack of reported data.
  • The thorough determination of the mass transfer coefficient and associated Sherwood number as a function of other non-dimensional parameters like gas holdup, bubble/slug lengths, and capillary number is left for future studies.
  • The study highlights the need for further investigation into the influence of bubble/slug lengths and other non-dimensional parameters on mass transfer.

Summary updated

BibTeX
@inproceedings{kuzmin2013lattice,
  author = {Alexander Kuzmin and Michał Januszewski and Dmitry Eskin and Farshid Mostowfi and J.J. Derksen},
  title = {{Lattice Boltzmann study of mass transfer for two-dimensional Bretherton/Taylor bubble train flow}},
  booktitle = {Chemical},
  year = {2013},
  doi = {10.1016/j.cej.2013.03.123},
}
Figure