Predictive modeling using adaptive PLS desulphurization reagent control system

Author

Dudzic, Michael ; Quinn, Shannon

Author_Institution

Dofasco Inc., Hamilton, Ont., Canada

Volume

1

fYear

2002

fDate

2002

Firstpage

604

Abstract

Projection to latent structures (partial least squares) is a multivariate statistical (MVS) technology that can be applied in inference/predictive control strategies. Dofasco has used MVS as the basis for an online, adaptive PLS prediction system, which is used to control its desulphurization facility. This paper introduces the reader to the basic process of steel desulphurization and motivation for the MVS modeling in this application. Also, the paper examines the PLS and associated adaptation algorithm used. Comments regarding online experiences and operational results are also given.

Keywords

adaptive control; chemical variables control; least squares approximations; predictive control; process control; statistical analysis; steel industry; Dofasco; adaptive control; chemical reactants; desulphurization; latent structure projection; multivariate statistical analysis; partial least squares; predictive control; steel industry; Adaptive control; Chemical processes; Control system synthesis; Control systems; Covariance matrix; Kernel; Memory; Predictive models; Programmable control; Steel;

fLanguage

English

Publisher

ieee

Conference_Titel

American Control Conference, 2002. Proceedings of the 2002

ISSN

0743-1619

Print_ISBN

0-7803-7298-0

Type

conf

DOI

10.1109/ACC.2002.1024874

Filename

1024874