Model Predictive Control of a Fed-batch Bioreactor Based on Dynamic Metabolic-Genetic Network Models

摘要

In this work, a model predictive control of a fed-batch bioreactor is presented, described by the dynamic enzyme-cost FBA model (deFBA). The deFBA model is employed within a bilevel optimization to obtain fed-batch operating policies including the substrate feeding and process-level regulation of metabolism for optimizing the productivity of a target product. The advantages of implementing the closed-loop control in order to compensate for modelling errors are evaluated by comparing with the performance of an open-loop control. A case study involving the fed-batch fermentationof Escherichia colifor ethanol production is considered to find optimal operating strategies for maximal productivity.