Real-time optimization of a fed-batch bioreactor with substrate inhibition using extremum-seeking

This paper proposes two real-time optimization strategies for single biomass fed-batch processes where it is desired to maximize a production or degradation rate (also called specific metabolic respiration rate), by estimating the corresponding optimal substrate concentration on the basis of basic measurements related to the feed input, the outlet gas composition (i.e., easily available measurements) and a limited knowledge of the model. A simple feedback controller only requiring the a priori knowledge of one coefficient is designed and tested in combination with the real-time optimization scheme, on a toxic wastewater treatment process.