(550i) Quantification of the Effect of Bubble Size Dynamics to Capture Interphase Mass Transfer and Gas Holdup in Bubble Column Reactors

Gas fermentation is a promising pathway for producing carbon based fuels and chemicals from captured CO2 and renewable feedstock gases such as green hydrogen. Optimizing the design of fermentation reactors (such as bubble columns, air-lift reactors) for application-specific gas fermentation and ensuring scale-up of the design requires predictive computational modeling approaches to elucidate bubble dynamics and mass transfer phenomena.

In bubble column reactors, the distribution of bubble sizes can influence critical design parameters such as interphase mass transfer rates and gas hold-up. Therefore, the hydrodynamics and the bubble size distribution (BSD) must be accurately predicted to scale-up and design bubble reactors [1]. For computational tractability, the BSD is best described by population-balance-based models that require closure models associated with bubble breakage and coalescence. In this work, the focus is on the simulation of bubble column reactors operating with CO2/CO/H2 mixtures . Using a constant bubble size assumption, our current simulation results obtained were on par with previous one-dimensional investigations [3,4] while consistently deviating from experimental measurements [2] (Fig. 1).

The central questions we address in this work are 1) “does adequately capturing bubble dynamics improve upon current model predictions with experiments?” and 2) “Are current population balance models applicable to CO2/CO/H2 gas mixtures where component gas species have highly varied solubilities and buoyancies?” We adopt a Bayesian inference approach where physics parameters are probabilistically inferred to fit experimental data. Statistical models are used to represent the dynamics of the bubble distribution and are adjusted via an efficiency factor to identify how to explain the discrepancy to experimental observations.

The multiphase flow is modeled using Reynolds Averaged Navier-Stokes (RANS) equations to represent mass and momentum transport, similar to [5]. Since forward runs (the 3D CFD runs) are expensive, we propose a computationally efficient approach to perform the Bayesian calibration.

[1]. Lo, S., and Zhang, D., Modelling of Break-up and Coalescence in Bubbly Two-Phase Flows. The Journal of Computational Multiphase Flows. 2009. 1: 23-38

[2] Deckwer, W‐D., I. Adler, and A. Zaidi. "A comprehensive study on co2‐interphase mass transfer in vertical cocurrent and countercurrent gas‐liquid flow." The Canadian Journal of Chemical Engineering 56.1 (1978): 43-55.

[3] Hissanaga, A. M., N. Padoin, and E. E. Paladino. "Mass transfer modeling and simulation of a transient homogeneous bubbly flow in a bubble column." Chemical Engineering Science 218 (2020): 115531.

[4] Ngu, V. and Morchain, J. and Cockx, A. Spatio-temporal 1D gas-liquid model for biological methanation in lab scale and industrial bubble column Chemical Engineering Science 251 (2022): 117478.

[5]. Rahimi, M., Sitaraman, H., Humbird, D. and Stickel, J., Computational fluid dynamics study of full-scale aerobic bioreactors: Evaluation of gas–liquid mass transfer, oxygen uptake, and dynamic oxygen distribution, Chemical Engineering Research and Design. 2018. 139: 283–295.