MODELING ERROR IDENTIFICATION OF ACTIVATED-SLUDGE MODELS

Information about the location of modeling errors is crucial for the e fficient improvement of an invalid model. This article discusses how t o pinpoint modeling errors through comparison of experimental data wit h data obtained through simulation of the invalid model. An observer-b ased approach is presented, By designing a dedicated observer for the system using the invalid model, a signal vector is generated, on which each modeling error imposes an easily identifiable feature. An algori thm to analyze the featured signal is then presented. With this algori thm, the features of each of the modeling errors are extracted, The ap proach is illustrated for a denitrification reactor model in which err ors in the dimension of the state vector, in the structure of the biok inetic relationship and in the values of the parameters could be ident ified. (C) 1997 IAWQ. Published by Elsevier Science Ltd.