Advanced nonlinear multi-layer process control for autotrophic cultivations ?

摘要

This paper presents a control architecture suitable for autotrophic cultivations in a stirred tank reactor feeding a gas mixture of H2/CO2/O2. Based on the employed multi-layer structure, for simple gas phase control no metabolic information of the strain is required. A MIMO-PI controller connected to an adaptive feedforward disturbance rejection based on a physical model allows for a gas phase feedback control. Thus, without further knowledge of the autotrophic strain, experiments can be run to generate data and develop a process model including strain specific biological details. Combining both layers, gas phase control and feeding trajectory optimization with the process model, results in a control architecture that is suitable for optimizing the process in a closed-loop.