Multirate state and parameter estimation in an antibiotic fermentation with delayed measurements.

Area of Science:

• Biotechnology
• Biochemical Engineering
• Process Monitoring

Background:

• Traditional software sensors like the extended Kalman filter (EKF) struggle with fermentation processes exhibiting endogenous metabolism and time-varying maintenance.
• These culture-related activities complicate accurate estimation and monitoring of critical fermentation parameters.

Purpose of the Study:

• To develop a robust multirate software sensor-based estimation strategy for monitoring fermentation processes.
• To address the limitations of traditional algorithms in handling complex metabolic activities.
• To accurately estimate biomass and specific growth rates in fed-batch fermentations.

Main Methods:

• Lumped endogenous decay and maintenance into a modified maintenance coefficient, m(c).
• Developed model equations relating measurable outputs (CER, biomass) to observable parameters (net specific growth rate, modified maintenance coefficient).
• Implemented a multirate estimation strategy accommodating delayed biomass measurements and frequent CER measurements.

Main Results:

• Successfully monitored biomass profiles and critical fermentation parameters, including specific growth rate.
• Demonstrated the effectiveness of the modified maintenance coefficient approach.
• Validated the multirate software sensor strategy in a fed-batch fermentation of Streptomyces clavuligerus.

Conclusions:

• The proposed multirate software sensor strategy effectively monitors fermentation processes with endogenous metabolism and time-varying maintenance.
• The modified maintenance coefficient approach enhances the accuracy of biomass and growth rate estimations.
• This method offers a reliable solution for complex fermentation monitoring challenges.