OPTIMAL EXPERIMENTAL-DESIGN FOR STRUCTURE CHARACTERIZATION OF BIODEGRADATION MODELS - ONLINE IMPLEMENTATION IN A RESPIROGRAPHIC BIOSENSOR

The interaction between activated sludge and a wastewater is subject t o important changes. This is reflected not only in changing biokinetic parameters but also in changing model structures. The need to select the `right' model structure in a reliable way on the basis of respirog raphic data provided by on-line sensors imposes serious real-time cons traints on the methods used. First of all, fast structure characteriza tion methods are presented allowing us to track the model structure on -line. Since these so-called a priori methods are less computing inten sive, they carl be at the basis of optimal experimental design calcula tions that can be performed on-line, This allows us to maintain the qu ality of the overall model identification under the changing process c onditions of a wastewater treatment plant. Two applications of Optimal Experimental Design are given.