DOI: dx.doi.org/10.21440/2307-2091-2016-2-66-69
Usage of neural network model for configuring of automatic gas turbine power regulators pdf
G. A. Kilin, I. R. Ziyatdinov, B. V. Kavalerov
The article discusses the use of neural network mathematical model for the automated setup of automatic control system of a gas turbine unit (GTU), consisting of a gas turbine power plant. The peculiarity of the construction of models based on neural network technology lies in the fact that the structural and parametric identification of the object takes place automatically during the process of training the neural network. Thus, we greatly simplify the technology and reduce the time of receipt of mathematical models for a GTU and GTU-SG system. A unique feature of the developed complex of techniques consists in the fact that they allow you to automatically, without human intervention, obtain mathematical models GTUs and promptly implement configured (tested) on a mathematical model, automatic control systems (ACS) on the real objects. The usage of software system significantly reduces the cost of developing, testing, ACS GTU, as all the experiments on the installation of ACS GTU are carried out on the resultant GTU model, rather than the real object, and this leads to fuel, human and time resources savings. As a result, improving the ACS GTU for power plants improves the main indicators of quality of generated electricity. However, it is important to make a follow-up inspection of the solutions on the full-scale facility before industrial using of the obtained adjustment regulators of GTU. In the long term, on the basis of the algorithm, it is possible to implement a mobile setup complex (MSC), with which the expert, having set an experiment on a real installation and downloading the results in MSC, receives a mathematical model of GTU, at which he can configure ACS GTU like it is really operating.
Keywords: neural network; mathematical model; identification; automatic control system; gas turbine unit; automated setup; regulator.
REFERENCES
Polulyakh A. I., Lisovin I. G., Kavalerov B. V., Shigapov A. A. 2011, Avtomatizatsiya nastroyki regulyatorov gazoturbinnykh mini-elektrostantsiy pri
komp’yuternykh ispytaniyakh [Automation of setting of regulators gas turbine power plants during a computer tests]. Avtomatizatsiya v promyshlennosti [Automation in industry], no. 6, pp. 14–17.
Kilin G. A., Kavalerov B. V., Bakhirev I. V. 2015, Poluchenie nelineynoy modeli GTU na osnove neyronnoy seti [Getting a nonlinear GTU model based on neural
network]. Avtomatizatsiya v elektroenergetike i elektrotekhnike [Automation in power industry and electrical engineering], no. 1, pp. 72–77.
Chobenko V. N., Palienko R. V., Lyutikov A. L. 2013, Matematicheskaya model’ odnoval’nogo GTD D045 [Mathematical model of single-shaft gas turbine engine
D045]. Vostochno-Evropeyskiy zhurnal peredovykh tekhnologiy [Eastern-European Journal of Enterprise Technologies], vol. 3, no. 63, pp. 18–21.
Kilin G. A. 2015, Poluchenie nelineynoy matematicheskoy modeli sistemy «gazoturbinnaya ustanovka–sinkhronnyy generator» s pomoshch’yu identifikatsii
[Getting nonlinear mathematical “gas turbine plant – synchronous generator” model system using identification]. Vestnik Izhevskogo gosudarstvennogo tekhnicheskogo universiteta [Bulletin of the Izhevsk State Technical University], no. 2, pp. 87–91.
Asgari H. et al. 2013, Artificial neural network-based system identification for a single-shaft gas turbine. Journal of Engineering for Gas Turbines and Power, vol.
135, no. 9. Available at: http://gasturbinespower.asmedigitalcollection.asme.org
Asgari H., Chen X. Q., Sainudiin R. 2013, Modeling and simulation of gas turbines. International Journal of Modeling, Identification and Control, vol. 20, no.
3, pp. 253–270.
Khizhnyakov Yu. N. 2014, Mnogosvyaznoe adaptivnoe upravlenie parallel’noy rabotoy sinkhronnykh generatorov s primeneniem neyronnoy tekhnologii [Multicoupling adaptive control of parallel operation of synchronous generators using neural technology], Moscow, vol. 16, 19 p.
Sil’verstov A. N., Chinaev P. I. 1987, Identifikatsiya i optimizatsiya avtomaticheskikh sistem [Identification and Optimization of automatic systems], Moscow,
200 p.
Khaykin S. 2008, Neyronnye seti: polnyy kurs [Neural networks: a complete course], Moscow, 1104 p.
Kruglov V. V., Borisov V. V. 2001, Iskusstvennye neyronnye seti. Teoriya i praktika [Artificial neural networks. Theory and practice], Moscow, 382 p.
Teyl G. 1971, Ekonometricheskie prognozy i prinyatie resheniy [Econometric forecasting and decision-making], Moscow, 488 p.
