Roth, Felix; (2025) Experimentally driven predictive soft sensors and economic models for continuous upstream bioprocesses. Doctoral thesis (Ph.D), UCL (University College London).

Text Roth_10217487_thesis.pdf Access restricted to UCL open access staff until 1 December 2026. Download (5MB)

Abstract

Continuous upstream processing using perfusion cell culture has regained significant traction in the biopharmaceutical industry due to its potential advantages in volumetric productivity, facility flexibility, and product quality. However, hurdles remain for widespread adoption as high viable cell density (VCD) cultures, enabled by advances in cell line development and media optimisation, have brought perfusion systems close to their physical limits in terms of mass transfer, filter flux, and nutrient supply. To manage these factors and ensure stable, long-term operation of high volumetric productivity cultures, novel process control mechanisms and design approaches are required. This project aimed to address this through the development of soft sensors and predictive models derived from multivariate analysis of on-line process data, in particular, oxygen transfer, pressure, viscosity and filter fouling effects. Soft sensors measuring viable cell density and filter fouling in real time were developed. These were integrated with process simulations, and economics and uncertainty analyses were conducted in order to provide insights into the impact of these factors on manufacturing processes. To validate the model, hypothetical production scenarios looking at different process durations and scales were compared in terms of cost, space-time yield, and robustness. Trade-offs between these factors were quantified, allowing more informed, data-driven process design choices. Together, these tools help overcome both perceived and real barriers to adoption of perfusion processes in industry as they advance process understanding and set a development framework that can be used for any continuous upstream process. The research presented in this work highlights the potential for novel process models and soft sensors to be integrated as core parts in the development and operation of modern, high cell density perfusion processes. The utility of these methods in commercial settings was demonstrated and can potentially offer significant contributions to process understanding for both industry and regulatory bodies.

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as CSVJSONXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item