Researchers have demonstrated a smart manufacturing framework for lab-scale continuous downstream processing of monoclonal antibodies, using modular automation hardware, interoperable software, and model-based control strategies to improve process monitoring and flow stability.
The work focused on a legacy bench-top platform for continuous precipitation-based capture purification of monoclonal antibodies. The process was developed as an alternative to Protein A chromatography and combines precipitation, dewatering, and countercurrent washing steps using static mixers and hollow fiber microfiltration modules.
Continuous downstream processing can offer efficiency and scale-up advantages, but it is difficult to automate at lab scale. In this system, one of the main challenges was flow control. Peristaltic pumps are widely used in bioprocess development because they are inexpensive and reduce contamination risk, but they can suffer from pulsation, drift, and calibration uncertainty. Those fluctuations become more problematic when several pumps, membranes, and recycle streams are connected in a continuous process.
To address these issues, the researchers retrofitted the platform with a smart manufacturing architecture built around RedLion FlexEdge edge devices, OPC-UA communication, Python and MATLAB control scripts, and cloud connectivity through the Smart Manufacturing Interoperability Platform. The system linked multi-vendor equipment including pumps, weigh-scales, pressure sensors, turbidity sensors, and an optical particle imaging probe.
The team then tested estimation and control methods using water-based experiments before moving to more complex monoclonal antibody streams. A Kalman filter was used to generate real-time flowrate estimates from noisy scale measurements. Steady-state data reconciliation was used to enforce material balance across measured and unmeasured flow streams. Proportional–integral controllers were used to adjust pump speeds and maintain target flowrates.
In closed-loop tests, the controllers tracked setpoints across four pumps and stabilized flowrates after step changes. Some interaction between pump control loops was observed, particularly in the interconnected filtration section. Adding a ratio constraint within the data reconciliation framework reduced those interactions and improved flow stability.
The study also showed how legacy lab equipment can be integrated into a more flexible automation architecture. A previous PendoTech system had limited input and output capacity, but the new modular setup could accommodate additional sensors, pumps, and scales while preserving compatibility with existing hardware.
The authors argue that this kind of approach could lower the barrier for process development labs to adopt automation, especially where budget constraints, mixed-vendor systems, and legacy equipment make full-scale digital transformation difficult.
The work remains a lab-scale demonstration, and the pump-control experiments were performed with water rather than monoclonal antibody process streams. Future work will need to test the system under biologically relevant conditions, where protein aggregation, membrane fouling, and concentration effects add further complexity.
Newsletters
Receive the latest pharmaceutical news, personalities, education, and career development – weekly to your inbox.
