Simulation results for on-line optimization of a batch bioreactor using nonlinear filtering and optimal control

Area of Science:

• Biotechnology
• Chemical Engineering
• Process Control

Background:

• Optimal control profiles for batch bioreactors often rely on simplified, empirical models.
• Model uncertainties can significantly impact the reliability of calculated control profiles.
• Existing methods may not adequately adapt to real-time fermentation dynamics.

Purpose of the Study:

• To develop a robust method for calculating near-optimal control trajectories in batch bioreactors.
• To address the challenge of model uncertainty in bioreactor process control.
• To create an adaptive control strategy for running fermentations.

Main Methods:

• Implemented a moving time horizon approach for successive control profile calculations.
• Utilized a mathematical model with kinetic parameters estimated by a nonlinear observer.
• Combined a nonlinear estimator with an optimizer for adaptive control.

Main Results:

• Numerical simulations were performed on xanthan-gum batch fermentations.
• The proposed method demonstrated reasonably good results in generating adaptive control trajectories.
• The nonlinear estimator plus optimizer arrangement proved effective for batch fermentors.

Conclusions:

• The developed method provides a reliable way to compute near-optimal control profiles for batch bioreactors.
• Adaptive control using real-time parameter estimation enhances fermentation process performance.
• This approach offers a promising advancement for optimizing bioprocesses with inherent model uncertainties.