Multi-Controller System: Major Boost to Continuous CHO Perfusion

Researchers at Sartorius have developed a nonlinear model predictive control system that significantly boosts continuous CHO cell perfusion. By simultaneously managing feed, bleed, and harvest flows, the system increased production by 68% while maintaining viable cell density above 95%. The multi-strategy NMPC system allows biomanufacturers to switch between stable operation and economic optimization seamlessly.

Dr. Mahshad Valipour and Christopher McCready, from Sartorius, explain that the system accounts for unknown states through a moving horizontal estimate, improving numerical conditioning and avoiding instability issues. By determining optimal operating conditions in real-time, the ENMPC mode maximizes performance objectives while enforcing dynamic process feasibility and biological constraints.

The primary advantage of this approach is the ability to determine maximum feasible operating conditions safely in real time, considering the accumulation of inhibitory biomaterials that hinder cell productivity. Conventional models fail to account for the trade-offs among perfusion rate, inhibitor removal, and cell growth, making them unsuitable for maximizing viable cell density.

Experiments using digital twins demonstrated the NMPC's ability to transition smoothly between process controller strategies, maintaining a healthy cell culture and increasing upstream productivity. Real-world evaluations confirmed that online and inline measurements of viable and dead cell density matched predictions. Notably, the system can change objectives during runtime, even when facing a significant plant-model mismatch.

This study suggests that advanced automation strategies can resolve the challenges of continuous CHO perfusion, making it more predictable and efficient.