Parameter and state estimation of the activated sludge process—I. Model development

A new approach is introduced for combined state and parameter estimation, named Linearized Maximum Likelihood (LML). The LML method is motivated by the uncertain and nonlinear activated sludge dynamics. In this article, the theoretical background leading to the development of the LML method is presented, along with the resulting numerical algorithm. In a companion article (Kabouris and Georgakakos, 1996) the LML method is applied to the parameter and state estimation of a nitrifying activated sludge process, modelled by the IAWPRC Activated Sludge Model 1.