PWA dynamic identification for nonlinear model fault detection

This chapter addresses the identification of nonlinear dynamic systems. A wide class of these systems can be described using nonlinear time-invariant regression models that can be approximated by means of piecewise affine prototypes with an arbitrary degree of accuracy. This chapter concerns the identification of piecewise affine model structure through input-output data acquired from a dynamic process. The identification approach exploited to estimate order and parameters of the local affine models is based on the Frisch scheme method for Error-in-Variables (EIV) models. An optimization technique, taking into account both continuity constraint fulfillment and EIV noise assumptions, is considered. To show theeffectiveness of the developed technique, when exploited also for Fault Detection and Isolation (FDI) purpose, the results obtained in the identification of both a simple simulated system and a real dynamic process are reported. © 2007 Copyright © 2007 Elsevier Ltd All rights reserved.