and Computational Methods

doi: 10.18698/2309-3684-2015-2-7386

In the paper we propose an algorithm of parameters (time constants of the turbine) iden-tification using the gradient method with an adaptive model. The adaptive mathematical model has the same structure as the identification object. The identification criterion is based on the loss function, which is the misalignment between the left and right sides of the equation, which describes the adaptive model. Thus it is avoided the need of finding the solution of a nonlinear equation for the adaptive model in an explicit form. In the model the signal observed at the output of the identified object is used instead of the output signal. Since mathematical models are nonlinear, the Newton – Kantorovich linearization and the matrix operator apparatus are applied to solve the problem. The features of gradient vector computation and features of the identification algorithm and its organization are considered. The results of the two time constants identification for the mathematical model of the turbine PT-12/15-35/10M are presented.

[1] Melnikov D.V., Fisher M.R. Vestnik MGTU im. N.E. Baumana. Seria Mashinostroenie – Herald of the Bauman Moscow State Technical University. Series: Mechanical Engineering, 2011, spetsialnyi vypusk “Energeticheskoe i transportnoe machinostroenie” [special issue “Power and transport mechanical engineering”], pp. 197–215.

[2] Kornyushin Yu.P., Melnikov D.V., Egupov N.D., Kornyushin P.Yu. Vestnic MGTU im. N.E. Baumana. Seria Estestvennye nauki – Herald of the Bauman Moscow State Technical University. Series: Natural Sciences, 2014, no. 1, pp. 78–93.

[3] Melnikov D.V., Egupov N.D. Izvestiya Tulskogo gosudarstvennogo universiteta. Tekhnicheskie nauki – Proceedings of the Tula State University. Engineering sciences, 2011, issue. 5, part 1, pp. 108–113.

[4] Kalashnikov A.A. Dinamika regulirovaniya turbin [Turbine control dynamics]. Moscow, Energoatomizdat Publ., 1999, 328 p.

[5] Tsypkin Ya.Z. Osnovy informatsionnoy teorii identificatsii [Basics of information theory of identification]. Moscow, Nauka Publ., 1984, 320 p.

[6] Gayskiy V.A., Egupov N.D., Kornyushin Yu.P. Primenenie funktsiy Walsha v sistemakh avtomatizatsii nauchnykh issledovaniy [Application of the Walsh functions in research automation]. Kiev, Naukova Dumka Publ., 1993, 212 p.

[7] Pupkov K.A., Egupov N.D., eds. Metody inzhenernogo sinteza slozhnykh system upravleniya: analiticheskiy apparat, algoritmy prilozheniya v tekhnike. Chast II. Vychislitelno-analiticheskiy eksperiment: apparat matrichnykh operatorov i vychislitelnye tekhnologii [Methods for the engineering synthesis of complex control systems: an analytical apparatus, algorithms of applications in engineering. Part II. Computational and analytical experiment: The matrix operators apparatus and the computational technology]. Moscow, BMSTU Publ., 2012, 416 p.

Kornyushin Y., Egupov N., Kornyushin P. Identification of parameters of regulator actuators for steam power turbines using matrix operator apparatus. Маthematical Modeling and Coтputational Methods, 2015, №2 (6), pp. 73-86

Количество скачиваний: 374