Другие журналы

scientific edition of Bauman MSTU

SCIENCE & EDUCATION

Bauman Moscow State Technical University.   El № FS 77 - 48211.   ISSN 1994-0408

Identification of Nonlinear Dynamic Systems Possessing Some Non-linearities

# 07, July 2015
DOI: 10.7463/0715.0789774
Article file: SE-BMSTU...o234.pdf (1025.00Kb)
authors: Y.N. Pavlov, V.M. Nedashkovskii, E.A. Tihomirova

The subject of this work is the problem of identification of nonlinear dynamic systems based on the experimental data obtained by applying test signals to the system. The goal is to determinate coefficients of differential equations of systems by experimental frequency hodographs and separate similar, but different, in essence, forces: dissipative forces with the square of the first derivative in the motion equations and dissipative force from the action of dry friction. There was a proposal to use the harmonic linearization method to approximate each of the nonlinearity of "quadratic friction" and "dry friction" by linear friction with the appropriate harmonic linearization coefficient.
Assume that a frequency transfer function of the identified system has a known form. Assume as well that there are disturbances while obtaining frequency characteristics of the real-world system. As a result, the points of experimentally obtained hodograph move randomly. Searching for solution of the identification problem was in the hodograph class, specified by the system model, which has the form of the frequency transfer function the same as the form of the frequency transfer function of the system identified. Minimizing a proximity criterion (measure) of the experimentally obtained system hodograph and the system hodograph model for all the experimental points described and previously published by one of the authors allowed searching for the unknown coefficients of the frequency transfer function of the system model. Solution of nonlinear dynamic system identification in frequency hodograph was reduced to solving the system of equations, which is linear relative to the unknown parameters of the frequency transfer function of the system model. The paper shows the possibility to identify a nonlinear dynamic system with multiple nonlinearities, obtained on the experimental samples of the frequency system hodograph. The proposed algorithm allows to select the nonlinearity of the type "quadratic friction" and "dry friction", i.e. also in the case where the nonlinearity is dependent on the same dynamic parameter, in particular, on the derivative of the system output value. For the dynamic system of the second-order with nonlinearity of the type "quadratic friction" in combination with nonlinearity of the type "dry friction", was developed a software to simulate a process for  providing pseudo experimental data containing random accuracy and to determine the parameters of the system. A conducted computational experiment enabled an estimate of the accuracy with which the proposed algorithm determines the parameters of the system. The illustrative numerical simulation has demonstrated that with using the proposed nonlinear dynamic system identification algorithm in frequency hodograph the accuracy of determining the coefficient values of the frequency transfer function of the second order system with a dry and quadratic friction is comparable with the range of measurement accuracy of experimental samples of this system hodograph. Well-known publications do not mention this identification method of the nonlinear dynamic systems. The nonlinear dynamical systems identification method the article describes can find application when determining parameters of various kinds of actuators. The using method of harmonic linearization and identification of dynamical systems by hodographs is promising for solving the problem of the identification of nonlinear systems with different types of nonlinearities.

References
  1. Zhdanov A.I., Katsiuba O.A. Identification by the method of least squares of autoregression equation parameters with additive measurement errors. Avtomatika i telemekhanika , 1982, no. 2, pp. 29-38. (English translation of journal : Automation and Remote Control , 1982, vol. 43, no. 2, pp. 158-166.).
  2. Bolkvadze G.R. Hammerstein Models for Identification of Stochastic Systems. Avtomatika I telemekhanika , 2003, no. 1, pp. 42-55. (English translation of journal : Automation and Remote Control , 2003, vol. 64, iss. 1, pp. 37-48. DOI: 10.1023/A:1021820307806 ).
  3. Pavlenko S.V. Application of wavelet filtering in the process of identification of nonlinear systems in the form of kernels of Volterra. Vostochno–evropeiskii zhurnal peredovykh tekhnologii = Eastern-European Journal of Enterprise Technologies , 2010, vol. 6, no. 4 (48), pp. 65-70. (in Russian).
  4. Raibman N.S., Kapitonenko V.V., Ovsepian F.A., Varlaki P.M. Dispersionnaia identifikatsiia [Dispersion identification]. Moscow, Nauka Publ., 1981. 336 p. (in Russian).
  5. Pavlenko S.V. Identification of Nonlinear Dynamic Systems in the Form of the Volterra Kernels on the Basis of the Data of Pulse Response Measurements. Elektronnoe modelirovanie , 2010, vol. 32, no. 3, pp. 3-18. (in Russian).
  6. Rajbman N.S. Extensions to nonlinear and minimax approaches. In: Eykhoff P., ed. Trends and Progress in System Identification . Oxford, Pergamon Press, 1981, pp. 185-237. (Russ ed.: Rajbman N.S. Metody nelineinoi i minimaksnoi identifikatsii. In: Eykhoff P., ed. Sovremennye metody identifikatsii sistem: sb. Moscow, Mir Publ., 1983, pp. 177-277.).
  7. Popov E.P., Pal'tov I.P. Priblizhennye metody issledovaniya nelineynykh avtomaticheskikh system [Approximate methods of study of nonlinear automatic systems]. Moscow, GIFML Publ., 1960. 790 p. (in Russian).
  8. Pugachev V.S., ed. Osnovy avtomaticheskogo upravleniya [Basics of automatic control]. Moscow, Nauka Publ., 1968. 680 p. (in Russian).
  9. Boevkin V.I., Pavlov Yu.N. Regressionnyy analiz v prikladnoy zadache identifikatsii [Regression analysis in applied problem of identification]. Moscow, Bauman MSTU Publ., 1990. ( Trudy MGTU im. N. E. Baumana № 546 [Proceedings of the Bauman MSTU; no. 546]). (in Russian).
  10. Boevkin V.I., Nedashkovskii V.M., Pavlov Iu.N. Identification of linear dynamic elements using a frequency locus. Nauka i obrazovanie MGTU im. N.E. Baumana = Science and Education of the Bauman MSTU , 2013, no. 9, pp. 349-360. DOI:10.7463/0913.0618917 (in Russian).
  11. Pavlov Yu.N., Nedashkovskii V.M, Tikhomirova E.A., Shavyrin I.B. Harmonic linearization method in the identification of nonlinear dynamical systems. Nauka i obrazovanie MGTU im. N.E. Baumana = Science and Education of the Bauman MSTU , 2014, no. 4, pp. 382-397. DOI: 10.7463/0414.0704613 (in Russian).

 

Поделиться:
 
SEARCH
 
elibrary crossref ulrichsweb neicon rusycon
Photos
 
Events
 
News



Authors
Press-releases
Library
Conferences
About Project
Rambler's Top100
Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00)
  RSS
© 2003-2024 «Наука и образование»
Перепечатка материалов журнала без согласования с редакцией запрещена
 Phone: +7 (915) 336-07-65 (строго: среда; пятница c 11-00 до 17-00)