EXPLORATORY ANALYSIS OF BIOPROCESSES USING ARTIFICIAL NEURAL-NETWORK-BASED METHODS

A process data driven procedure has been developed that allows a unive rsal time-efficient bioprocess analysis. The procedure is particularly suited for industrial production processes which have not yet been co mprehensively investigated. It makes use of artificial neural networks in combination with mass balance equations to represent the process d ynamics on a commercial workstation. The essential concept behind the procedure is to start with the already available knowledge formulated by a very simple process representation which includes only those vari ables that are firmly known to be essential. Then, stepwise, additiona l variables are added to the basic representation after they passed a test procedure in which they proved to enhance the model's performance . The result of the procedure is a numerical representation of the imp ortant process relationships that immediately allows to determine impr oved set points and/or profiles for the manipulated variables with res pect to process performance. It may be used to improve state estimatio n and control. The procedure has already been tested in industrial app lications. In this paper, a validation of the procedure with simulated bioprocess data is presented.