Modelling large-scale bubble column bioreactors using computational fluid dynamics

Scale-up of any bioprocess to full industrial scale is a considerable technical challenge, typically related to an increased degree of flow inhomogeneity that leads to the formation of undesirable gradients (for example, of dissolved oxygen or substrate). Being able to quantify these large-scale gradients, and hence identify ways in which their impacts can be minimized, is of obvious interest in the design, operation and optimisation of large-scale fermenters. Working in collaboration with a global manufacturing partner, we have developed a Computational Fluid Dynamics (CFD) model of bubble columns operating at the high superficial velocities commonly required in commercial systems. In this presentation, we discuss the challenges involved in developing our model and extending it from the bench-top (15 L) through to the pilot (120 L) and commercial (80,000 L) scales. As part of this model development program, an extensive experimental campaign was undertaken at all scales to provide the necessary validation and test data. Our latest version currently runs on two commercial CFD platforms, each of which can predict observed experimental behaviour to within engineering accuracy (10-20%) for all operating conditions and scales evaluated. In addition, our presentation will discuss the inclusion of microbial kinetics into the model together with the use of a particle tracking system that allows internal inhomogeneities to be quantified in terms of their impact on commercial operational parameters such as substrate to product yield. The presentation concludes with an overview of how this model is being used in a range of ongoing commercial studies.