Machine learning driven quantification of local bubble dynamics across a pilot scale stirred bioreactor

Chemical Engineering Science, vol. 339, pp. 125011

Abstract

Understanding gas bubble behavior is critical for oxygen transfer and bioprocess performance, yet localized characterization under non-ideal operating conditions remains limited. This study introduces a methodology combining inline shadowgraphy with convolutional neural networks (CNNs) to achieve spatial analysis of gas bubbles in a stirred pilot scale reactor. The approach quantifies bubble size, gas holdup, and specific interfacial area across varying aeration rates, stirrer speeds, and liquid viscosities. Results highlight the importance of inline shadowgraphy for capturing bubble dynamics in complex media. Viscosity was found to dominate over inertial forces, producing smaller bubbles and disrupting flow patterns at higher values. This can lead to bubble trapping in low shear zones, increasing gas holdup and specific interfacial area. Operating parameters were also shown to influence bubble count and distribution, with clear implications for local homogeneity and the availability of specific interfacial area. The methodology provides a versatile, open-source, and transferable pipeline for local bubble characterization, offering practical guidance for process optimization and scale-up in gas-dependent bioprocesses, including aerobic fermentation, gas-based fermentation, and oxidative biocatalysis.

Authors 6

1. Emilie Overgaard Willer

   Technical University of Denmark

   Affiliation as printed

   Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark

2. Maximilian Siska Aachen Computational Systems Biotechnology

   Forschungszentrum Jülich · RWTH Aachen University

   Affiliation as printed

   Computational Systems Biotechnology, RWTH Aachen University, Forckenbeckstraße 51, 52074 Aachen, Germany

   Institute of Bio- and Geosciences, Forschungszentrum Jülich GmbH, Wilhelm-Johnen-Straße, 52428 Jülich, Germany

3. Benjamin Petersen

   Technical University of Denmark

   Affiliation as printed

   Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark

4. Juliet J. Victoria

   Technical University of Denmark

   Affiliation as printed

   Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark

5. Eric von Lieres corresponding Aachen Computational Systems Biotechnology

   Forschungszentrum Jülich · RWTH Aachen University

   Affiliation as printed

   Computational Systems Biotechnology, RWTH Aachen University, Forckenbeckstraße 51, 52074 Aachen, Germany

   Institute of Bio- and Geosciences, Forschungszentrum Jülich GmbH, Wilhelm-Johnen-Straße, 52428 Jülich, Germany

6. John M. Woodley corresponding

   Technical University of Denmark

   Affiliation as printed

   Chemical and Biochemical Engineering, Technical University of Denmark, 2800 Kgs. Lyngby, Denmark