Hybrid Approach to State Estimation for Bioprocess Control.

Area of Science:

• Biotechnology
• Chemical Engineering
• Control Systems

Background:

• Accurate state estimation is critical for optimizing bioprocess control.
• Traditional methods struggle with complex, data-driven observation models.
• Recombinant therapeutic protein production requires precise monitoring of key variables.

Purpose of the Study:

• To develop an improved state estimation technique for bioprocess control.
• To enhance the Unscented Kalman Filter (UKF) with a hybrid approach.
• To enable estimation of the specific biomass growth rate (µ) for advanced control.

Main Methods:

• A hybrid Unscented Kalman Filter (UKF) was developed.
• The dynamic system model used ordinary differential equations for mass balances.
• An observation model was formulated using Support Vector Regression (SVR).
• The method was applied to a recombinant therapeutic protein production process in Escherichia coli.

Main Results:

• The hybrid UKF demonstrated effective state estimation for bioprocesses.
• The specific biomass growth rate (µ) was successfully estimated.
• State estimates exhibited a low noise level, suitable for advanced control.

Conclusions:

• The proposed hybrid UKF offers a robust solution for bioprocess state estimation.
• Accurate estimation of biomass growth rate facilitates innovative control strategies.
• This technique is well-suited for advanced bioprocess control applications, particularly in therapeutic protein production.