Abstract

In order to monitor and control bioprocesses, online measurements of certain biological variables, like biomass concentration, are normally required. Unfortunately, there is a lack of highly reliable and low cost instruments for online measuring these important variables. These limitations have motivated the use of observers. This paper describes a state observer for bioprocesses based on a hybrid fuzzy model. The model combines the mass balances and a fuzzy inference system for approximating the uncertain dynamic of the kinetic reactions. The convergence, stability, and robustness of the observer are analyzed. The main features of this observer are illustrated by considering a lactic fermentation. © 2004 IEEE.