(394h) The Composer: Lessons Learned on Developing a Data-Driven Assistant for Bioreactor Design and Optimization

The impact of heterogeneities in the environment experienced by micro-organisms can lead to uncertainties in reactor performance upon scale-up. Computational Fluid Dynamics (CFD) simulations with integrated biokinetics have emerged as a tool to assess this impact [1]. Nevertheless, its inherent computational demand hampers a fast and reliable evaluation of reactor operation schemes. As an alternative, compartment models (CM) have shown the capability to capture hydrodynamic features in stirred tank bioreactors at an affordable computational cost [2]. This is true for both steady conditions [3] and dynamic conditions, such as in fed-batch fermentations [4, 5, 6].

Leveraging the benefits of transformers [7] and compartment modelling [4, 5, 8] as a means for creating surrogate CFD-based models of bioreactors has enabled flexibility in modelling bioreactor performance upon changes in operating conditions. This has unveiled new possibilities for assisting in the exploration of bioreactor geometries and operating schemes. Thus, we share insight on the challenges faced on the path towards a generative-AI workflow for bioreactor design and optimization.

References

[1] C. Haringa, W. Tang, G. Wang, A. T. Deshmukh, W. A. van Winden, J. Chu, W. M. van Gulik, J. J. Heijnen, R. F. Mudde and H. J. Noorman, "Computational fluid dynamics simulation of an industrial P. chrysogenum fermentation with a coupled 9-pool metabolic model: Towards rational scale-down and design optimization," Chemical Engineering Science, vol. 175, pp. 12-24, 2018.

[2] A. Delafosse, "CFD-based compartment model for description of mixing in bioreactors," Chemical Engineering Science, vol. 106, pp. 76-85, 2014.

[3] T. Tajsoleiman, R. Spann, C. Bach, K. V. Gernaey, J. K. Huusom and U. Krühne, "A CFD based automatic method for compartment model development," Computers & Chemical Engineering, vol. 123, pp. 236-245, 2019.

[4] G. Nadal-Rey, D. D. McClure, J. M. Kavanagh, B. Cassells, S. Cornelissen, D. F. Fletcher and K. V. Gernaey, "Development of dynamic compartment models for industrial aerobic fed-batch fermentation processes," Chemical Engineering Journal, vol. 420, p. 130402, 2021.

[5] J. L. De Carfort, T. Pinto and U. Krühne, "An automatic method for generation of CFD-Based 3D compartment models: Towards Real-Time Mixing Simulations," Bioengineering, vol. 11, p. 169, 2024.

[6] H. Maldonado de Leon, A. Straathof and C. Haringa, "Dynamic compartment models: Towards a rapid modeling approach for fed-batch fermentations," Chemical Engineering Science, vol. 308, pp. 1-16, 2025.

[7] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser and I. Polosukhin, "Attention is All You Need," in 31st Conference on Neural Information Processing Systems, Long Beach, 2017.

[8] C. Haringa, W. Tang and H. J. Noorman, "Stochastic parcel tracking in an Euler–Lagrange compartment model for fast simulation of fermentation processes," Biotechnology and Bioengineering, vol. 119, no. 7, pp. 1849-1860, 2022.