Lancaster EPrints

Application of evolutionary learning in Wiener neural identification and predictive control of a plug-flow tubular reactor

Arefi, Mohammad Mehdi and Montazeri, Allahyar and Jahed-Motlagh, Mohanimad Reza and Poshtan, Javad (2007) Application of evolutionary learning in Wiener neural identification and predictive control of a plug-flow tubular reactor. In: Industrial Electronics Society, 2007. IECON 2007. 33rd Annual Conference of the IEEE. IEEE Industrial Electronics Society . IEEE, New York, pp. 644-650. ISBN 978-1-4244-0783-5

Full text not available from this repository.

Abstract

In this paper, identification and nonlinear model predictive control of highly nonlinear plug-flow tubular reactor based on Wiener model is studied. This process simulated in a rather realistic environment by HYSYS, and the obtained data is in connection with MATLAB for identification and control purpose. The process is identified with NN-Wiener identification method, and two linear and nonlinear model predictive controllers are applied with the ability of rejecting slowly varying unmeasured disturbance. Since the identification problem must be solved with a nonlinear optimization method, to attain the best possible model for prediction genetic algorithm is used. The Simulation results show that the obtained Wiener model has a good capability to predict the step response of the process. The results for control are also compared with a common PI controller for temperature control of tubular reactor. It is shown that the nonlinear controller has the fastest damped response in comparison with the other two controllers.

Item Type: Contribution in Book/Report/Proceedings
Uncontrolled Keywords: SYSTEMS ; STATE ; INTERNAL MODEL CONTROL
Subjects:
Departments: Faculty of Science and Technology > Engineering
ID Code: 65701
Deposited By: ep_importer_pure
Deposited On: 15 Jul 2013 11:01
Refereed?: Yes
Published?: Published
Last Modified: 21 Sep 2017 02:37
Identification Number:
URI: http://eprints.lancs.ac.uk/id/eprint/65701

Actions (login required)

View Item