0
TECHNICAL PAPERS: Gas Turbines: Structures and Dynamics

Neural Network Emulation of a Magnetically Suspended Rotor

[+] Author and Article Information
A. Escalante, V. Guzmán, M. Parada, L. Medina, S. E. Diaz

Universidad Simon Bolivar, Decanato de Investigacion y Desarrollo, Caracas 1080-A, Venezuela

J. Eng. Gas Turbines Power 126(2), 373-384 (Jun 07, 2004) (12 pages) doi:10.1115/1.1689363 History: Received December 01, 2001; Revised March 01, 2002; Online June 07, 2004
Copyright © 2004 by ASME
Your Session has timed out. Please sign back in to continue.

References

Bhat,  N. V., Minderman,  P. A., McAvoy,  T., and Wang,  N. S., 1990, “Modeling Chemical process System Via Neural Computation,” IEEE Control Syst. Mag., Apr.
Chen,  S., Billings,  S. A., and Grant,  P. M., 1990, “Non-Linear Systems Identification using Neural Networks,” Int. J. Control, 51, pp. 1191–1214.
Nguyen,  D., and Widrow,  B., 1990, “Neural Networks for Self-Learning Control Systems,” IEEE Control Syst. Mag., Apr.
Chen, H., and Lewis, P., 1992, “Rule-Based Damping Control for Magnetic Bearings,” Proc. 3rd International Symposium on Magnetic Bearings, Technomic, Lancaster, PA, pp. 25–32.
Fittro, R., 1993, “Magnetic Bearing Control Using Artificial Neural Networks,” MAG’93 Magnetic Bearings, Magnetic Drives and Dry Gas Seals Conference, Alexandria, VA, Technomic, Lancaster, PA, pp. 201–210.
Choi, B. B., Brown, G., and Johnson, D., 1997, “Neural Network Control of a magnetically Suspended Rotor System,” MAG’97, Magnetic Bearings, Magnetic Drives and Dry Gas Seals Conference, Alexandria, VA, Technomic, Lancaster, PA, pp. 281–289.
Billings,  S. A., Jamaludin,  H. B., and Chen,  S., 1992, “Properties of Neural Networks With Application to Modeling Non-linear Dynamical System,” Int. J. Control, 55, pp. 194–224.
Nahi, N. E., 1969, Estimation Theory and Applications, Krieger, New York.
Goodwing, G. C., and Payne, R. L., 1977, Dynamic System Identification: Experiment Design and Data Analysis, Academic Press, San Diego.
Ljung, L., and Soderstrom, T., 1983, Theory and Practice of Recursive Identification, M.I.T. Press, Cambridge, MA.
Narendra,  K. S., and Parthasarathy,  K., 1990, “Identification and Control of Dynamical System Using Neural Networks,” IEEE Trans. Neural Netw., 1, pp. 4–27.
Levenberg,  K., 1944, “A Method for Solution of Certain Nonlinear Problems in Least Squares,” Q. Appl. Math., 2, pp. 164–168.
Marquardt,  D., 1963, “An Algorithm for Least Squares Estimation of Nonlinear Parameters,” SIAM (Soc. Ind. Appl. Math.) J. Appl. Math., 11(2), pp. 164–168.
Hagan,  M. T., and Menhaj,  J. B., 1994, “Training Feedforward Networks With the Marquardt Algorithm,” IEEE Trans. Neural Netw., 5(6), pp. 989–993.
Norgaard, M., Ravn, O., Poulsen, N. K., and Hansen, L. K., 2000, Neural Networks for Modeling and Control of Dynamic System, 1st Ed., Springer-Verlag, Berlin.
Haykin, S., 1998, Neural Networks: A Comprehensive Foundation, 2nd Ed., Prentice-Hall, Englewood Cliffs, NJ.
Ljung, L., 1999, System Identification-Theory for the User, 2nd Ed., Prentice-Hall, Englewood Cliffs, NJ.
Margaglio, E., 1998, “Recurrent Neural Networks Applied to Non-linear System Identification,” Internal Report, Simón Bolı́var University (in Spanish).
Billings,  S. A., and Zhu,  Q. M., 1994, “Nonlinear Model Validation Using Correlation Test,” Int. J. Control, 60(6), pp. 1107–1120.
Sweitzer, G., Bleuler, H., Traxler, A., 1994, Active Magnetic Bearings: Basics, Properties, and Applications of Actives Magnetic Bearing, 1st Ed., v/d/|f Hochschulverlag AG an der ETH Zurich, Zurich.

Figures

Grahic Jump Location
Node or artificial neuron
Grahic Jump Location
The multilayer perceptron
Grahic Jump Location
System Identification procedure
Grahic Jump Location
Magnetic bearing working principle
Grahic Jump Location
Experimental setup lock diagram
Grahic Jump Location
w and v-axis identification
Grahic Jump Location
Single axis control scheme
Grahic Jump Location
Ramp No. 2 training sequence (normalized amplitude versus number of samples)
Grahic Jump Location
Validation sequence, 2000 samples per channel, with “double unbalance” and running at 3000 rpm. (3000 rpm2d) (normalized amplitude versus number of samples).
Grahic Jump Location
Artificial neural network structure
Grahic Jump Location
Autocorrelation functions, ANN trained with Ramp1 and tested with 3000 rpm validation sequence (correlations values versus number of delayed samples)
Grahic Jump Location
Autocorrelation functions, ANN trained with Ramp2 and tested with 3000 rpm validation sequence (correlations values versus number of delayed samples)
Grahic Jump Location
Autocorrelation functions, ANN trained with Ramp3 and tested with 3000 rpm validation sequence (correlations values versus number of delayed samples)
Grahic Jump Location
Autocorrelation function, ANN trained with Ramp4 and tested with 3000 rpm validation sequence (correlations values versus number of delayed samples)
Grahic Jump Location
TCV prediction error ANN trained with Ramp2 using the Levenberg-Marquardt method (currents values in A. versus number of samples).
Grahic Jump Location
TCV correlation functions. ANN trained with Ramp2, using the Levenberg-Marquardt method (correlations values versus number of delayed samples).
Grahic Jump Location
TCV output. ANN trained with Ramp2, using the back-propagation algorithm (currents values in A. versus number of samples).
Grahic Jump Location
TCV correlation functions. ANN trained with Ramp2, using the back-propagation algorithm (correlations values versus number of delayed samples).
Grahic Jump Location
TCW output, testing done with 3000 rpm sequence (currents values in A. versus number of samples)
Grahic Jump Location
Correlation functions, TCW output, testing done with 3000 rpm sequence (correlations values versus number of delayed samples)
Grahic Jump Location
TCW output, testing done with 3000 rpmd sequence (currents values in A. versus number of samples)
Grahic Jump Location
Correlation functions, TCW output, testing done with 3000 rpmd sequence (correlations values versus number of delayed samples)
Grahic Jump Location
TCW output, testing done with 3000 rpm2d sequence (currents values in A. versus number of samples)
Grahic Jump Location
Correlation functions, TCW output, testing done with 3000 rpm2d sequence (correlations values versus number of delayed samples)

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In