Soft Sensor Based Control of Complex Biochemical Processes

Abstract Fed-batch production processes for recombinant proteins are of the prime importance in modern biotechnology. In order to run the high performance process, it is necessary to keep it in optimal path. Nevertheless, at present there is no reliable methods for the on-line determination of key process variables and to use estimation results for control purposes. It is shown that soft-sensors for the estimation of μ, and the concentrations of biomass X, inducer L and protein Ρ can be of considerable advantage to keep these processes under control. They can be based on model supported measurements and empirical correlations. Their high efficiency and accuracy were demonstrated experimentally.