Abstract
On line identification of nonlinear causal relationships between variables in the feed water control loop of a nuclear power plant is reported. The knowledge about the observable variables of the application has been used in the design of the architecture for the network, in the local function of each elemental processor (quadratic expansion of inputs and recurrence) and, finally, in the selection of the supervised learning algorithm. This learning algorithm is based on the local evaluation and propagation of individual output errors for each sample in the training set. This nonlinear model with delays and quadratic expansion of inputs is compared with the more usual linear dynamic network and a clear improvement is observed. Some preliminary conclusions on the influence of signal noise relationships and the criteria for the selection of the appropriated sampling period are also included.
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
Barto, G. A. (1990): “Conectionist Learning for Control”. Neural Networks for Control. Miller III, Thomas W.; Sutton, Richard S.; Werbos, Paul J. (Eds.) pp. 5–58. The MIT Press.
Cellier, F. E. (1991): “Continuous System Modeling”, pp. 659–671. Springer-Verlag.
Fonseca, J.S. da; McCulloch, W.S. (1967): “Synthesis and Linearization of Non-Linear Feedback Shift Registers: Basis of a Model of Memory”. Q.P.R. n∘ 86. pp. 355. MIT.
Fukunaga, K. (1964): “A Theory of Non-Linear Autonomous Sequential Nets Using Z-Transformes”. IEEE Trans., 1964, vol. EC-13, pp. 310–313.
Gallant, S.I. (1993): “Neural Network Learning and Expert Systems”. The MIT Press.
General Electric: “Design Specification” (Nuclenor). Plant documentation of Santa Maria de Garoña Nuclear Power Plant.
Mira, J. (1992): “Computación Neuronal”. UNED. Madrid.
Mira, J.; Delgado, A.E.(1991): “Always Trying to Write an Equation for the Brain”, Lecture Notes in Computer Science, 540, Artificial Neural Networks. A. Prieto (de.) pp 93–100. Springer-Verlag.
Mira, J.; Delgado, A.E.(1991): “Linear and Algorithmic Formulation of Co-operative Computation in Neural Nets.” Lecture Notes in Computer Science, 585, Computer Aided Systems Theory, EuroCast'91. F. Pichler and R. Moreno Diaz (Eds.) pp 2–20. Springer-Verlag.
Mira, J.; Delgado, A.E.; Álvarez, J.R.; P. de Madrid, A.; Santos, M. (1993a): “Towards More Realistic Self Contained Models of Neurons: High-Order, Recurrence and Local Learning” New Trends in Neural Computation, pp. 55–62. J.Mira, J.Cabestany, A.Prieto (Eds.) Springer-Verlag.
Mira, J.; Fernández, R.A.; Álvarez, J.R. (1993b): “Evaluación, Análisis y Modelación de los Sistemas de Control de la Central Nuclear de Santa Maria de Garoña”. (Project Report) Madrid y Santander.
Mira, J.; Delgado, A.E.; Santos, M.; P. de Madrid, A.; Álvarez, J.R. (1994): “Local Learning in Networks of Universal Analogic Neurons”. Cybernetics and Systems 25: pp. 259–273.
Narendra, K.S. (1990): “Adaptive control usign neural networks”. Neural Networks for Control. Miller III, Thomas W.; Sutton, Richard S.; Werbos, Paul J. (Eds.) pp. 5–58. The MIT Press.
Nilsson, N. (1965): “Learning Machines”.pp. 27–30. McGraw-Hill.
Lettvin, J.Y. (1962): “Form-function Relations in Neurons”, Q.P.Q. n∘ 66, pp. 333. Research Lab. of Elect. MIT.
Röckmann, D.; Moraga, C. (1991): “Using Quadratic Perceptrons to Reduce Interconnection Density in Multilayer Neural Networks”. Lecture Notes in Computer Science, 540, Artificial Neural Networks, A. Prieto (de.) pp. 86–91, Springer-Verlag.
Rumelhart, D.E.; Hinton, G.E.; Williams, R.J. (1986): “Learning internal representations by error propagation”, Parallel Distributed Processing: Explorations in the Microstructure of Cognition, Vol 1 Foundations. D.E. Rumelhart and J.L. McClelland (Eds.). The MIT Press.
Werbos, P.J. (1988): “Generalization of back propagation with applications to a recurrent gas market model” Neural Networks, 1. pp 339–356.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1995 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Álvarez, J.R., Mira, J., Fernández, R.A., Sainz, L., Arroyo, V., Delgado, A.E. (1995). On line identification of causal relationships between variables in the feed water system of a nuclear power plant. In: Mira, J., Sandoval, F. (eds) From Natural to Artificial Neural Computation. IWANN 1995. Lecture Notes in Computer Science, vol 930. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-59497-3_282
Download citation
DOI: https://doi.org/10.1007/3-540-59497-3_282
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-59497-0
Online ISBN: 978-3-540-49288-7
eBook Packages: Springer Book Archive