State and Parameters Estimation by Extended Kalman Filter for Studying Inhomogeneous Dynamics in Industrial Bioreactors

摘要

A software sensor (SS) for state and parameter estimation is proposed. The designed SS algorithm presents a modification on extended Kalman filter (EKF) for class biotechnological processes. The state vector is extended with two vectors of kinetic parameters - reaction rates and yield coefficients. Experimental data of fed-batch cultivation of a non-sporulating B. subtilis mutant realized in two compartment reactor are used as data basis. The measurements of dry cell weight and substrate concentration are considered as SS input information. The results show that applying new EKF algorithm the process monitoring is extended by main kinetic parameters. In such a way studying of inhomogeneous process dynamics could be improved.