004.89 Parametric identification of aerodynamic characteristics of a transport category aircraft using recurrent semi-empirical neural networks in the Tensorflow environment
Kreerenko S. S. (PJSC Beriev Aircraft Company), Kreerenko O. D. (PJSC Beriev Aircraft Company)
DYNAMICS, AIRPLANE, IDENTIFICATION, SEMI-EMPIRICAL, NEURAL NETWORK, RECURRENT, TENSORFLOW, KERAS, TRAINING
doi: 10.18698/2309-3684-2024-3-8199
The problem of modeling the longitudinal motion of a transport category aircraft and the parametric identification of the aerodynamic characteristics of the longitudinal motion: the components of the dimensionless coefficients of aerodynamic lift and pitching moment are considered. The problem is solved in a class of modular semiempirical dynamic models created by combining theoretical and neural network modeling. The performance and practical significance of the models is confirmed by the results of computational experiments. The development of a neural network model of the longitudinal movement of an aircraft was carried out in Python using the Tensorflow open software library for machine learning and the high-level Keras API as part of Tensorflow.
[1] Vasilchenko K.K., Leonov V.A., Pashkovsky I.M., Poplavsky B.K. Letnoe ispytaniya samoletov [Flight testing of aircraft]. Moscow, Mashinostroenie Publ., 1996, 719 p.
[2] Vasilchenko K.K., Kochetkov Yu.A., Leonov V.A., Poplavsky B.K. Strukturnaya identifikaciya matematicheskoj modeli dvizheniya samoleta [Structural identification of a mathematical model of aircraft movement]. Moscow, Mashinostroenie Publ., 1993, 352 p.
[3] Linnik Yu.V. Metod naimen'shih kvadratov i osnovy matematiko-statisticheskoj teorii obrabotki nablyudenij [The method of least squares and the basics of the mathematical and statistical theory of observation processing]. Moscow, Fizmatgiz Publ., 1962, 333 p.
[4] Ljung L. System identification: theory for the user. USA, Prentice Hall, 1987, 530 p.
[5] Moiseev N.N., Ivanilov Yu.P., Stolyarova E.M. Metody optimizacii [Optimization methods]. Moscow, Nauka Publ., 1978, 352 p.
[6] Eykhoff P. System identification. London, Wiley, 1974, 555 p.
[7] Kulifeev Yu.B. Diskretno-nepreryvnyj metod identifikacii nepreryvnyh sistem [Discrete-continuous method of identification of continuous systems]. Mechanics of Solids, 1981, no. 5, с. 47–55.
[8] Belotserkovsky A.S., Kachanov B.O., Kulifeev Yu.B., Morozov V.I. Sozdanie i primenenie matematicheskih modelej samoletov [Creation and application of mathematical models of airplanes. Moscow], Nauka Publ., 1984, 143 p.
[9] Dreyfus G. Neural networks : methodology and applications. Berlin, New York, Springer, 2005, 497 p.
[10] Dreyfus G., Idan Y. The Canonical Form of Nonlinear Discrete-Time Models. Neural Computation, 1998, vol. 10, no. 1, pp. 133–164. DOI: 10.1162/089976698300017926
[11] Nerrand O., Roussel-Ragot P., Personnaz L., Dreyfus G., Marcos S. Neural networks and nonlinear adaptive filtering: unifying concepts and new algorithms. Neural Computation, 1993, vol. 5, no. 2, pp. 165–199. DOI: 10.1162/neco.1993.5.2.165
[12] Oussar Y., Dreyfus G. How to be a gray box: dynamic semi-physical modeling. Neural Computation, 2001, vol. 14, no. 9, pp. 1161–1172. DOI: 10.1016/S0893-6080(01)00096-X
[13] Egorchev M.V., Kozlov D.S., Tiumentsev Yu.V. Aircraft short-period longitudinal motion modelling: a comparison of theoretical, empirical and semi-empirical approaches. Civil Aviation High Technologies, 2015, no. 211 (1), pp. 116–123.
[14] Egorchev M.V., Tyumentsev Yu.V. Nejrosetevye poluempiricheskie modeli upravlyaemyh dinamicheskih sistem [Neural network semi-empirical models of controlled dynamic systems]. Sbornik nauchnyh trudov XV Vserossijskoj nauchno-tekhnicheskoj konferencii «Nejroinformatika-2013», Ch. 2 [Collection of scientific papers of the XV All-Russian Scientific and Technical Conference "Neuroinformatics-2013", Part 2], Moscow, MEPhI Publ., 2013, pp. 22–31.
[15] Egorchev M.V., Kozlov D.S., Tiumentsev Yu.V., Chernyshev A.V. Neural network based semi-empirical models for controlled dynamical systems. Vestnik komp'iuternykh i informatsionnykh tekhnologii (Herald of computer and information technologies), 2013, no. 9 (111), pp. 3–10.
[16] Egorchev M.V., Tyumentsev Yu.V. Obuchenie poluempiricheskoj nejrosetevoj modeli upravlyaemogo dvizheniya samoleta [Training of a semi-empirical neural network model of controlled aircraft movement]. Sbornik nauchnyh trudov XVI Vserossijskoj nauchno-tekhnicheskoj konferencii «Nejroinformatika-2014», Ch. 2 [Collection of scientific papers of the XV All-Russian Scientific and Technical Conference "Neuroinformatics-2014", Part 2], Moscow, MEPhI Publ., 2014, pp. 363–272.
Крееренко С.С., Крееренко О.Д. Моделирование и параметрическая идентификация аэродинамических характеристик самолета транспортной категории с использованием нейросетей в среде Тensorflow. Математическое моделирование и численные методы, 2024, № 3, с. 81–99.
Исследование выполнено без финансирования со стороны каких-либо организаций.
Download article
Количество скачиваний: 376