Simple nonlinear observers for on-line estimation of kinetic rates in bioreactors

In this paper, we propose a global theoretical framework for the analysis of bioreactor dynamics. We concentrate on the problem of kinetic rates' estimation inside the reactor. We show how efficient estimators of these kinetics can be designed in a systematic and rigorous way from a minimal knowledge of their structures. The proposed approach does not require any model for the kinetic rates and is very successful in accurately estimating these variables. It uses technical tools issued from the non-linear systems theory, which allow the design of observer-based estimators. These estimators are designed to work with continuous and discrete data measurements. The continuous and discrete versions of the corresponding observers are presented and their convergences are established. An important point lies in the facility of implementing and calibrating the proposed estimators. Indeed, due to the particular structure of their gain, their tuning is reduced to the calibration of a single parameter whatever the number of considered components and reactions. The performances of the proposed estimators are illustrated in simulation through a typical example dealing with microbial growth and biosynthesis reactions. (C) 1998 Elsevier Science Ltd. All rights reserved.