Model predictive control made accessible to professional automation systems in fermentation technology

The objective of this paper is showing how model predictive control can easily be implemented directly into industrial bioreactor automation systems and thus making this control technology accessible to the fed-batch fermentation processes.By means of a practical example it is shown how to keep the biomass concentration exactly on its predefined path taking the substrate feed rate as the only action variable in a bioreactor that is conventionally equipped with standard measurement devices.The model predictive control algorithm uses a very simple general process model the parameters of which are adapted during the cultivation process.Additionally the feed rate profile corresponding to the desired biomass profile is used as a scheduling variable to adapt to changes in the process dynamics and at the same time for safeguarding