Neural network modeling and nonlinear predictive control of a biotechnological fed-batch process

Fed-batch fermentation processes are common methods of producing biological recombinant from different microorganisms. Model-based control of bioprocesses is a difficult task due to the challenges associated with bioprocess modeling. The paper deals with the multilayer perceptron neural network modeling of fed-batch cultivation of E. coli BL21 (DE3) [pET3a-ifnγ] under maximum attainable specific growth rate in the whole of process for producing γ-Interferon protein based on experimental data. The neural network model based predictive control (NNMPC) scheme is designed for controlling the concidered fed-batch biotechnological process. The nonlinear predictive control based on this model show good performance for tracking reference trajectories.