Generate QR code
Open Access
Subscription Access
Development of NPP steam turbine neural network model
https://doi.org/10.24223/1999-5555-2025-18-2-112-121
Abstract
The results are presented of a study of the possibility of using neural network (NNT) technologies to analyze the energy efficiency of power plants through the development of an NS model of a steam turbine installation (ST) of a VVER-1000/320 design NPP. The development of such models is quite a challenge, since it involves solving many tasks: determining the target functions — the parametrs that the NNT model will determine; searching for input parameters and extracting a variable part from them using various statistical analysis methods; choosing the type of NNT model experimentally; developing the NNT as such and the software package based on it, which would be accessible and understandable to the personnel of the power plant. During the study, all the presented measures were carried out, and a software package was developed to determine the values of the target functions, namely, specific heat consumption (SHC), specific steam consumption (SSC), electrical efficiency and the instantaneous value of the specific consumption of conventional fuel (SCCF), based on eighty-two input parameters, of which two are determined as variable ones. At the same time, the accuracy in validating the NNT model on a sample that did not participate in the NNT training is 99.5 %. Therefore, the developed NNT model makes it possible to determine the necessary indicators with high accuracy, with the time spent on calculating the same being within 0.5 s. This confirms the possibility of using NNT models to evaluate the energy efficiency of power plants according to the algorithm presented in the study, and opens up prospects for optimizing the operating mode of power plant equipment using similar NNT models. In order to demonstrate the capabilities of the software package, technically sound standards for the relevant target functions have been developed on its basis, which can be recommended for use in analyzing the energy efficiency of the ST under consideration.
About the Authors
P. A. Mineev
Ivanovo State Power University
Russian Federation
Russian Federation
Department NPP
153003; Rabfakovskaya str., 34; Ivanovo
V. A. Gorbunov
Ivanovo State Power University
Russian Federation
Russian Federation
153003; Rabfakovskaya str., 34; Ivanovo
N. A. Lonshakov
Ivanovo State Power University
Russian Federation
Russian Federation
153003; Rabfakovskaya str., 34; Ivanovo
M. N. Mechtaeva
Ivanovo State Power University
Russian Federation
Russian Federation
153003; Rabfakovskaya str., 34; Ivanovo
References
1. Rybakov D. A. Relevance and accessibility of neural networks in modern society. Mezhdunarodny`j nauchny`j zhurnal «Vestnik Nauki» 2023; (7): 256 – 261. (In Russ.)
2. Komlichenko V. N., Fedosenko V. A., Kupreichik A. S. Comparative analysis of various neural network architectures for regression problems. E`konomika i kachestvo sistem svyazi 2025; (1): 110 – 121. (In Russ.)
3. Intelligent decision support system for the management of turboprop pumps at Kalinin NPP. V. A. Gorbunov, N. A. Lonshakov, S. A. Teplyakova, M. N. Mechtaeva, P. A. Mineev. Izvestiya vuzov: yadernaya energetika 2024; (3): 125 – 140. (In Russ.)
4. The Russian Federation. Laws. On Energy Conservation and Energy Efficiency Improvement and on Amendments to Certain Legislative Acts of the Russian Federation : Federal Law No. 261-FZ: [adopted by the State Duma on November 11, 2009]. 2009: 52. (In Russ.)
5. Methodology for assessing the effect of operational parameters on the operation of a turbopump pump. V. A Gorbunov, N. A. Lonshakov, S. A. Teplyakova, M. N. Mechtaeva, P. A. Mineev. Vestnik IGE`U. 2022; (4): 14 – 23. (In Russ.)
6. GOST R 57700.23-2020. Computer models and modeling. Validation. General provisions : National Standard of the Russian Federation : date of introduction November 13, 2020 Federal`noe gosudarstvennoe unitarnoe predpriyatie «Rossijskij federal`ny`j yaderny`j centr. Vserossijskij nauchno-issledovatel`skij institut eksperimental`noj fiziki». Standartinform 2020: 4. (In Russ.)
7. Application of factor analysis techniques in data warehouses for plan factor analysis. A. A. Proschenkov, I. Yu. Vashchilo, E. A. Barabanov, T. R. Gazdiev, E. V. Koptenok. Sovremennaya nauka, obshhestvo i obrazovanie 2022; (1): 93 – 96. (In Russ.)
8. Zhuikova E. G. Investigation of various methods of cluster analysis based on their applications in neural networks. Perspektivy nauki 2024; (6): 87 – 91. (In Russ.)
9. Feofilov D. S. Comparative analysis of artificial neural network training methods. Texnika XXI veka glazami molody`x ucheny`x i specialistov. 2022; (20): 382 – 385. (In Russ.)
10. Petrov I. B., Stankevich A. S., Vasyukov A. V. On the search for an initial approximation in the wave inversion problem using convolutional neural networks. Doklady` RAN: matematika, informatika, processy` upravleniya 2023; (1): 58 – 64. (In Russ.)
11. Ilyichov D. E., Golovushkin A. A., Junkin D. A. Neural network training methods: gradient descent, Newton's method, conjugate gradient. Nauka i inovacii v XXI veke 2021; (1): 55 – 57. (In Russ.)
12. Microsoft Visual Studio: official website. URL: https://visualstudio.microsoft.com/ru / (accessed 02/01/2025).
Review
For citations:
Mineev P.A., Gorbunov V.A., Lonshakov N.A., Mechtaeva M.N. Development of NPP steam turbine neural network model. Safety and Reliability of Power Industry. 2025;18(2):112-121. (In Russ.) https://doi.org/10.24223/1999-5555-2025-18-2-112-121
Views:
37
Email this article 
