Semi-Mechanistic and Black-Box Modeling of an Enzymatic Conversion Process

Abstract

Abstract In this paper, a semi-mechanistic and a black-box modeling approach are applied to the modeling of an enzymatic conversion of Penicillin which takes place in a fed-batch reactor. Because of the nonlinear and time-varying character of the process, it is very difficult to construct a black-box model solely based on experimental data. Therefore, first a semi-mechanistic model is developed from the experimental data. This semi-mechanistic model is then used to generate process input-output sequences which can be used for the black-box modeling stage. Experiments show that both models give accurate process output predictions.