Goal-directed Design of Dynamic Experiments for Cybernetic Models of Bioreactors

Abstract

Abstract The distinctive feature of cybernetic models of bioreactors is their capacity to account for regulatory mechanisms in a cell metabolism by modeling the synthesis of enzymes and their activities. From a process engineering viewpoint, to guarantee its predictive capabilities regarding one or more process objective or goals (e.g. optimization, controllability, etc.), experimental data used to fit a cybernetic model parameters should be the most informative bearing in mind the adequacy of the resulting model to describe the specific objective of interest. To excite the most relevant metabolic modes in the cybernetic model, a dynamic experiment is optimally designed by accounting for the sensitivity of the chosen objective to operating conditions. The bioreactor feeding profile and sampling times are selected to maximize a global sensitivity index. As a case study, biomass production or fermentation to ethanol conversion in the fed-batch cultivation of Saccharomyces Cerevisiae are considered as alternative objectives.