Non-linear observability of activated sludge process models

The main contribution of this paper is to present a non-linear observability analysis method of Activated Sludge Models (ASM), which are used in many control applications. The objective is to reduce the unobservable ASM1 model to an observable one that can be used to implement advanced estimation algorithms. Local observability is achieved under certain operating conditions but failed at some points in the whole domain of definition. Furthermore, piece-wise observability rank test is also performed with three measurements and compared with non-linear observability. Simulation results are presented to demonstrate the proposed method. Copyright © 2005 IFA