An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

Marcin Mrugalski

doi:10.2478/amcs-2013-0013

.blurhash-client-img { display: none !important; }

An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

International Journal of Applied Mathematics and Computer Science

Volume 23 (2013): Issue 1 (March 2013)

By: Marcin Mrugalski

Open Access

|Mar 2013

Abstract

This paper presents an identification method of dynamic systems based on a group method of data handling approach. In particular, a new structure of the dynamic multi-input multi-output neuron in a state-space representation is proposed. Moreover, a new training algorithm of the neural network based on the unscented Kalman filter is presented. The final part of the work contains an illustrative example regarding the application of the proposed approach to robust fault detection of a tunnel furnace.

References

Back, A. and Tsoi, A. (1991). FIR and IIR synapses. A new neural network architectures for time series modelling, Neural Computation 3(3): 375-385.
Search in Google Scholar
Bernstein, D. (2005). Matrix Mathematics, Princeton University Press, Princeton, NJ.
Search in Google Scholar
Ding, S. (2008). Model-based Fault Diagnosis Techniques:Design Schemes, Algorithms, and Tools, Springer-Verlag, Berlin/Heidelberg.
Search in Google Scholar
Elman, J. (1990). Finding structure in time, Cognitive Science14(2): 179-211.10.1207/s15516709cog1402_1
Search in Google Scholar
Fasconi, P., Gori, M. and Soda, G. (1992). Local feedback multilayered networks, Neural Computation4(1): 120-130.10.1162/neco.1992.4.1.120
Search in Google Scholar
Gori, M., Bengio, Y. and De Mori, R. (1989). BPS: A learning algorithm for capturing the dynamic nature of speech, InternationalJoint Conference on Neural Networks, Washington,DC, USA, pp. 417-423.
Search in Google Scholar
Gupta, M., Liang, J. and Homma, N. (2003). Static and DynamicNeural Networks, John Wiley & Sons, Hoboken, NJ.
Search in Google Scholar
Haykin, S. (2001). Kalman Filtering and Neural Networks, John Wiley&Sons, New York, NY.10.1002/0471221546
Search in Google Scholar
Haykin, S. (2009). Neural Networks and Learning Machines, Prentice Hall, New York, NY.
Search in Google Scholar
Ichalal, D., Marx, B., Ragot, J. and Maquin, D. (2012). New fault tolerant control strategies for nonlinear Takagi-Sugeno systems, International Journal of AppliedMathematics and Computer Science 22(1): 197-210, DOI: 10.2478/v10006-012-0015-8.10.2478/v10006-012-0015-8
Search in Google Scholar
Isermann, R. (2005). Fault-Diagnosis Systems: An Introductionfrom Fault Detection to Fault Tolerance, Springer-Verlag, Heidelberg/Berlin.10.1007/3-540-30368-5_18
Search in Google Scholar
Ivakhnenko, A. and Mueller, J. (1995). Self-organization of nets of active neurons, System Analysis Modelling Simulation20(1-2): 93-106.
Search in Google Scholar
Jordan,M. and Bishop, C. (1997). Neural networks, in A. Tucker (Ed.), CRC Handbook of Computer Science, CRC Press, Boca Raton, FL, pp. 536-556.
Search in Google Scholar
Julier, S. and Uhlmann, J. (2004). Unscented filtering and nonlinear estimation, Proceedings of the IEEE92(3): 401-422.10.1109/JPROC.2003.823141
Search in Google Scholar
Korbicz, J. and Kościelny, J. (Eds.) (2010). Modeling, Diagnosticsand Process Control: Implementation in the DiaSterSystem, Springer-Verlag, Berlin.10.1007/978-3-642-16653-2
Search in Google Scholar
Korbicz, J., Kościelny, J., Kowalczuk, Z. and Cholewa,W. (Eds.) (2004). Fault Diagnosis. Models, Artificial Intelligence,Applications, Springer-Verlag, Berlin.10.1007/978-3-642-18615-8
Search in Google Scholar
Korbicz, J. and Mrugalski, M. (2008). Confidence estimation of GMDH neural networks and its application in fault detection system, International Journal of System Science39(8): 783-800.10.1080/00207720701847745
Search in Google Scholar
Lee, T. and Jiang, Z. (2006). On uniform global asymptotic stability of nonlinear discrete-time systems with applications, IEEE Transactions on Automatic Control51(10): 1644-1660.10.1109/TAC.2006.882770
Search in Google Scholar
Ljung, L. (1999). System Identification: Theory for the User, Prentice Hall PTR, Upper Saddle River, NJ.
Search in Google Scholar
Montes de Oca, S., Puig, V., Witczak, M. and Dziekan, Ł. (2012). Fault-tolerant control strategy for actuator faults using LPV techniques: Application to a two degree of freedom helicopter, International Journal of Applied Mathematicsand Computer Science 22(1): 161-171, DOI: 10.2478/v10006-012-0012-y.10.2478/v10006-012-0012-y
Search in Google Scholar
Mrugalski, M. and Korbicz, J. (2007). Least mean square vs. outer bounding ellipsoid algorithm in confidence estimation of the GMDH neural networks, in B.
Search in Google Scholar
Beliczynski, A. Dzielinski, M. Iwanowski, and B. Ribeiro (Eds.), Adaptive and Natural Computing Algorithms,Part 2, Lecture Notes in Computer Science, Vol. 4432, Physica-Verlag, Heidelberg, pp. 19-26.
Search in Google Scholar
Mrugalski, M. and Korbicz, J. (2011). GMDH neural networks, in B. Wilamowski and J. Irwin (Eds.), The Industrial ElectronicsHandbook, 2nd Edn., Vol. 5, CRC Press, Taylor Francis Group, Boca Raton, FL, pp. 8-1-8-21.10.1201/b10604-11
Search in Google Scholar
Mrugalski, M. and Witczak, M. (2012). State-space GMDH neural networks for actuator robust fault diagnosis, Advancesin Electrical and Computer Engineering 12(3): 65-72.10.4316/aece.2012.03010
Search in Google Scholar
Mrugalski, M., Arinton, E. and Korbicz, J. (2003). Dynamic GMDH type neural networks, in L. Rutkowski and J.10.1007/978-3-7908-1902-1_108
Search in Google Scholar
Kacprzyk (Eds.), Neural Networks and Soft Computing, Heidelberg, Physica-Verlag, pp. 698-703.
Search in Google Scholar
Mrugalski, M., Witczak, M. and Korbicz, J. (2008). Confidence estimation of the multi-layer perceptron and its application in fault detection systems, Engineering Applications of ArtificialIntelligence 21(6): 895-906.10.1016/j.engappai.2007.09.008
Search in Google Scholar
Niemann, H.H. (2012). A model-based approach to fault-tolerant control, International Journal of AppliedMathematics and Computer Science 22(1): 67-86, DOI: 10.2478/v10006-012-0005-x.10.2478/v10006-012-0005-x
Search in Google Scholar
Noura, H., Theilliol, D., Ponsart, J. and Chamseddine, A. (2009). Fault-tolerant Control Systems: Design and Practical Applications, Springer-Verlag, London.10.1007/978-1-84882-653-3
Search in Google Scholar
Palade, V., Bocaniala, C. and Jain, L. (2006). ComputationalIntelligence in Fault Diagnosis, Springer-Verlag, London.10.1007/978-1-84628-631-5
Search in Google Scholar
Pan, Y., Sung, S. and Lee, J. (2001). Data-based construction of feedback-corrected nonlinear prediction model using feedback neural networks, Control Engineering Practice9(8): 859-867.10.1016/S0967-0661(01)00050-8
Search in Google Scholar
Patan, K., Witczak, M. and Korbicz, J. (2008). Towards robustness in neural network based fault diagnosis, International Journal of Applied Mathematicsand Computer Science 18(4): 443-454, DOI: 10.2478/v10006-008-0039-2.10.2478/v10006-008-0039-2
Search in Google Scholar
Patton, R., Frank, P. and Clark, R. (2000). Non-linear SystemsIdentification. From Classical Approaches to Neural Networksand Fuzzy Models, Springer-Verlag, Berlin.
Search in Google Scholar
Teixeira, B., Torres, L., Aguirre, L. and Bernstein, D. (2010). On unscented Kalman filtering with state interval constraints, Journal of Process Control 20(1): 45-57.10.1016/j.jprocont.2009.10.007
Search in Google Scholar
Williams, R. and Zipser, D. (1989). A learning algorithm for continually running fully recurrent neural networks, NeuralComputation 1(2): 270-280.10.1162/neco.1989.1.2.270
Search in Google Scholar
Witczak, M. (2007). Modelling and Estimation Strategies forFault Diagnosis of Non-Linear Systems. From Analyticalto Soft Computing Approaches, Springer-Verlag, Berlin.
Search in Google Scholar
Witczak, M., Korbicz, J., Mrugalski, M. and Patton, R. (2006). A GMDH neural network based approach to robust fault detection and its application to solve the Damadics benchmark problem, Control Engineering Practice14(6): 671-683.10.1016/j.conengprac.2005.04.007
Search in Google Scholar
Witczak, M. and Prętki, P. (2007). Design of an extended unknown input observer with stochastic robustness techniques and evolutionary algorithms, International Journalof Control 80(5): 749-762.10.1080/00207170601178108
Search in Google Scholar
Zamarreño, J. M. and Vega, P. (1998). State space neural network. Properties and application, Neural Networks11(6): 1099-1112.
Search in Google Scholar