Adaptive control of nonlinear pressure tanks using grey box modeling with neural networks

• Autorzy: Ma Z. , Jutan A.
• Języki publikacji: EN

Abstrakty

EN

In this work, a nonlinear process is modeled in a so called 'grey box' format. This format divides the model into a known linear part and an unknown nonlinear part. The linear part is modeled using simple linear discrete dynamics whereas the nonlinear pan is modeled using a neural network with some weights. An adaptive controller is designed to incorporate the unknown nonlinear dynamics into the overall process dynamics. A smaller number of neural network weights can be used in this format so that the controller can be used adoptively for on-line control. The effectiveness of this approach is demonstrated using simulation and finally is used for real time control of a pressure tank process with excellent results.

Słowa kluczowe

EN

nonlinear system; adaptive control; neural networks; grey-box model; real-time process control system; pressure tank

PL

system nieliniowy; regulacja adaptacyjna; sieci neuronowe; model "szarej skrzynki"; sterowanie w czasie rzeczywistym; zbiornik ciśnieniowy

• Wydawca: Wydawnictwo Politechniki Łódzkiej
• Czasopismo: Journal of Applied Computer Science
• Rocznik: 2000
• Tom: Vol. 8, nr 2
• Strony: 45--63
• Opis fizyczny: Bibliogr. 16 poz.

Twórcy

autor

Ma Z.

• Departament of Chemical & Biochemical Engineering, The University of Western Ontario, London Ontario, N6A 5B8, Canada

autor

Jutan A.

• Departament of Chemical & Biochemical Engineering, The University of Western Ontario, London Ontario, N6A 5B8, Canada

Bibliografia

• [1] Bhat, N.V. and TJ. McAvoy, 'Use of Neural Nets for Dynamical Modeling and Control of Chemical Process Systems', Computers Chem. Eng. 14, 573-583,1990.
• [2] Bose, N. K. and Liang, P. 'Neural Network Fundamentals with Graphs', Algorithms and Applications, McGraw-Hill, Inc., New York, 1996.
• [3] U.Borisson,'Self-tuning regulator for a class of multivariable systems’, Automatica, Vol. 15, pp. 209-215, 1979.
• [4] [4] Chai, T. Y., An indirect Stochastic Adaptive Scheme with on-line Choice of Weightings Polynomials, IEEE Trans. Auto. Contr. AC-34,1990.
• [5] Chen, V. C. and Y. H. Pao,'Learning Control with Neural Networks', In Proc. of IEEE Int. Conf. on Robotics and Automation, Scttsdale, AZ. 1448-1453, 1989.
• [6] [6] Clarke D. W.and Gawthrop P. J.,'Self-tuning controller1, Proceedings of the IEE, Vol. 122, pp.929-934, 1975.
• [7] Cybenko G., 'Approximation by Superposition's of a Sigmoidal Function', Dept. Elec. Computer Eng., Univ. Illinois at Urbana-Champaign, Thch. Rep.856, 1988.
• [8] David J. H., C. Wen and G. C. Goodwin, 'Stability Analysis of Decentralized Robust Adaptive Control’, Proc. 27th IEEE Conf. on Decision and Control, 296- 300, 1988.
• [9] Funahashi K., 'On the Approximate Realization of Continuous Mappings by Neural Networks’, Vol.2, 183-192, 1989.
• [10] Goldberg, K. and B. A. Pearlmutter,'Using a Neural Networks to learn the dynamic of the CMU Direct-drive Arm 1Г, Report CMU-CS-88-160,Camegie-Mellon University, 1988.
• [11] Goodwin G. C.and K. S. Sin, 'Adaptive filtering prediction and control ' (Information and System Science Series), Englewood Cliffs, NJ: Prentice- Hall, 1984.
• [12] Goodwin G. C., P.J. Ramadge, and P.E. 'Caines, Discrete-time multivariable system’, IEEE Trans. Automat. Contr. Vol. AC-25, pp. 449-456, 1980.
• [13] Hecht-Nielsen R.,'Theory of the Back-Propagation Neural Networks', Proc.IEEE, Int. Conf. Neural Networks, 1-593-605, 1989.
• [14] Hoenik K.,M.Stinchcombe and H. White, 'Multilayer Feedforward Networks are Universal Approxmators', Neural Networks, Vol.2, 359-366, 1989
• [15] Yeung,D. Y. and G. A. Beckey, 'Using a Context-sensitive Learning Networks for Robot Arm Control'. Proc. of IEEE Int. Conf. on Robotics and Automation, Scottsdale, AZ, 1441-1447,1989.
• [16] Zhu Q.M., K. Warwick and J.L.Douce,'Adaptive General Prodictive Controller for Nonlinear System1,IEE, Proc. D, Vol. 138, 33-40,1991.