Combining Mechanistic Modeling and Raman Spectroscopy for Real-Time Monitoring of Fed-Batch Penicillin Production

A soft sensor that combines data from Raman spectroscopy and off-gas analyzers with a dynamic mechanistic bioprocess model was investigated for online monitoring of the physiological characteristics of Penicillium chrysogenum fed-batch cultivations. A systematic workflow based on nonlinear observability analysis was established for accelerating and improving the process of soft-sensor development. Using in situ Raman spectroscopy, it was possible to perform accurate and frequent measurements of the penicillin concentration in the bioreactor, which were combined with measurements of respiratory rates. Using a particle filter algorithm, the soft sensor allowed for the online estimation of the biomass concentration, the specific growth rate, and the specific penicillin production rate.