14th IFAC Symposium on System Identification, SYSID 2006

SYSID-2006 Paper Abstract

Close

Paper ThA4.3

Harnischmacher, Gerrit (RWTH Aachen Univ.), Kahrs, Olaf (RWTH Aachen Univ.), Marquardt, Wolfgang (RWTH Aachen Univ.)

Identification of a Fluid Catalytic Cracking Unit by Means of a New Multi-Variable Hammerstein Model

Scheduled for presentation during the Regular Session "Identification in Process Industry" (ThA4), Thursday, March 30, 2006, 11:10−11:30, Cummings Room

14th IFAC Symposium on System Identification, March 29 - 31, 2006, Newcastle, Australia

This information is tentative and subject to change. Compiled on September 24, 2018

Keywords Multivariable System Identification, Nonlinear System Identification, Grey Box Modelling

Abstract

Recently, a new formulation of a multi-variable Hammerstein model has been proposed, which preserves the concept of a nonlinear static element cascaded by an independent linear dynamic element known from scalar Hammerstein models. We apply this model to the identification of a simulated fluid catalytic cracking unit (FCC) demonstrating the need for representing input-directional dynamics as well as nonlinear coupling of input variables in block-oriented chemical process modeling. Because of the preserved independence of the nonlinear and linear elements the model allows for physically meaningful intermediate variables, which facilitates the incorporation of rigorous models as well as arbitrary approximations thereof. The superiority of two such approximations, namely sparse grid and neural network representations over the classical polynomial representation is also demonstrated.