This paper designs pH and conductivity control systems for a biomanufacturing platform that specializes in the production of biologically derived medicines on demand. Dynamic models and control strategies are presented for a buffer mixing unit that must produce high quality buffers for use within various production steps including a perfusion bioreactor and chromatography separation train. As salt addition does not affect the pH of the buffer, we separate the conductivity and pH control problems. The buffer conductivity is controlled using a PID controller tuned for robustness. The reaction-invariant method is used to overcome inherent nonlinearities in the titration curve such that good closed-loop performance can be obtained for any pH setpoint. Maximum a posteriori estimation is used to estimate parameters online in order to adapt the model in the presence of plant-model mismatch. Lastly, a dual mixer system is proposed to accurately estimate any bias in the inlet flowrates, which is a common fault in these types of systems. The effectiveness of this control strategy is demonstrated in a relevant case study.

• Massachusetts Institute of Technology (MIT) (MIT)

• Control

1. Conductivity
2. Control
3. Mathemetical model
4. Mixers
5. PD control
6. Production
7. Proteins
8. Research Article