Community Research and Development Information Service - CORDIS

Abstract

Predictive control systems basically consist of introducing nonlinear models to describe the nonlinearities and complexities of dynamic systems along the future. The computational time requirements of that control systems may be an obstacle in real time applications because at every sampling time a nonlinear optimization problem must be solved. In this paper, a new nonlinear predictive control approach is proposed. It makes use of a neural network (NN) to calculate the predictive control action. Thus, the problems related to the high computational effort are reduced. The predictive neural controller is applied to the real-time control of the heat transfer fluid temperature in a chemical bath reactor and compared against the PID controller installed in the plant.

Additional information

Authors: ISIS, ;SAIA ISPRA (IT),
Bibliographic Reference: Paper presented: 9th International Symposium on System Modelling Control, Institute of Computer Science, Technical University of Lodz (PO), 18-21st May, 1998
Availability: Available from Public Relations and Publications Unit, Ispra (IT)
Record Number: 199910814 / Last updated on: 1999-06-07
Category: PUBLICATION
Original language: en
Available languages: en