Abstract
An application of neural net plant models to nonlinear model predictive control is presented. Feedforward neural nets are trained with input / output data from a plant under conventional control to obtain a black-box-model of the plant. This model is used for the prediction task in an extended DMC scheme. Experimental results for a SISO-example (continuous neutralization reactor) and a MIMO-example (level control) show a significant improvement of the controller performance compared to the conventional controllers used for generating the training data.
Preview
Unable to display preview. Download preview PDF.
References
K. Hornik, M. Stincombe, H. White: Universal Approximation of an Unknown Mapping and Its Derivatives Using Multilayer Feedforward Networks. Neural Networks 3, 211–223, 1990.
D. E. Rumelhardt, G. E. Hinton, R. J. Williams: Learning Internal Representations by Error Propagation. Parallel Distributed Processing 1, 45–76, MIT Press, 1986.
T. Tollenaere: SuperSAB: Fast Adaptive Backpropagation with Good Scaling Properties. Neural Networks 3, 561–573, 1990.
T. Peterson, E. Hernandez, Y. Arkun, F. J. Schork: Nonlinear Predictive Control of a Semi Batch Polymerisation Reactor by an Extended DMC. Proc. 1989 American Control Conference, 1534–1539, 1989.
C. Cutler, B. Ramaker: Dynamic Matrix Control — A Computer Control Algorithm, Proc. 1980 Joint Automatic Control Conference.
H. Ranke: Modellbasierte prädiktive Regelung eines Neutralisationsreaktors mit Neuronalen Netzen, Diplomarbeit, Universität Dortmund, Fachbereich Chemietechnik. 1993.
U. Meier: Mathematische Modellierung und Entwurf eines nichtlinearen Reglers für einen kontinuierlich betriebenen Neutralisationsreaktor, Diplomarbeit, Universität Dortmund, Fachbereich Chemietechnik, 1993.
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 1995 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Draeger, A., Engell, S. (1995). Nonlinear Model Predictive Control Using Neural Net Plant Models. In: Berber, R. (eds) Methods of Model Based Process Control. NATO ASI Series, vol 293. Springer, Dordrecht. https://doi.org/10.1007/978-94-011-0135-6_23
Download citation
DOI: https://doi.org/10.1007/978-94-011-0135-6_23
Publisher Name: Springer, Dordrecht
Print ISBN: 978-94-010-4061-7
Online ISBN: 978-94-011-0135-6
eBook Packages: Springer Book Archive