«Automation and Informatization of the fuel and energy complex»

Application of neural networks for diagnostics and prediction of equipment technical state

UDC: 004.032.26
DOI: 10.33285/2782-604X-2023-7(600)-13-21

Authors:

¹ InfTech, Moscow, Russia
² Ufa University of Science and Technology, Ufa, Russia

Keywords: diagnostics of industrial equipment, neural networks, assessment of the units operation reliability, prediction of the technical state, machine learning

Annotation:

Improving the reliability and stability of equipment operation is a crucial and priority task for any industrial enterprise. With the development of new communication technologies related to the transfer of technological processes data and their parameters, including physical indicators of equipment operation, in real time, the solution to this problem required new approaches to identify faults and possible failures. There appeared a necessity of developing and applying intelligent algorithms to analyze and process large amounts of data in order to detect the abnormal work of industrial units. The article considers the possibility of using machine learning technologies, in particular neural networks, for diagnosing and predicting the technical state of equipment. Three main directions of the studied problem solution are discussed: classification of the current state of the unit, prediction of the technical characteristics or attributes of the equipment and recognition of an industrial machine behavior by clustering algorithms. The structure of the system for diagnosing and predicting the industrial equipment state of is presented. Computational experiments and analysis of the obtained results were carried out.

Bibliography:

1. Mukhutdinov A.R., Efimov M.G., Vakhidova Z.R. Neyrosetevoe modelirovanie protsessa detonatsii smesevogo vzryvchatogo veshchestva na osnove geksida // Vestn. Tekhnolog. un-ta. – 2020. – T. 23, № 1. – S. 84–88.
2. Shmelev V.A. Avtomatizirovannye sistemy upravleniya protsessom bureniya neftyanykh i gazovykh skvazhin, sostoyanie razrabotok // Avtomatizatsiya, telemekhanizatsiya i svyaz' v neftyanoy promyshlennosti. – 2021. – № 9(578). – S. 49–59. – DOI: 10.33285/0132-2222-2021-9(578)-49-59
3. Naletov V.A., Glebov M.B., Naletov A.Yu. Staticheskaya matematicheskaya model' nagrevatel'noy pechi dlya optimizatsii protsessov energopotrebleniya // Avtomatizatsiya i informatizatsiya TEK. – 2022. – № 8(589). – S. 50–56. – DOI: 10.33285/2782-604X-2022-8(589)-50-56
4. Convolutional LSTM network: A machine learning approach for precipitation nowcasting / Xingjian Shi, Zhourong Chen, Hao Wang [et al.] // Advances in neural information processing systems. – 2015. – P. 802–810. – DOI: 10.48550/arXiv.1506.04214
5. Peel L. Data driven prognostics using a Kalman filter ensemble of neural network models // 2008 Int. Conf. on prognostics and health management, Denver, CO, USA, Oct. 06–09, 2008. – IEEE, 2008. – 6 p. – DOI: 10.1109/PHM.2008.4711423
6. Prognostics and health management design for rotary machinery systems – Reviews, methodology and applications / Jay Lee, Fangji Wu, Wenyu Zhao [et al.] // Mechanical systems and signal processing. – 2014. – Vol. 42, Issue 1-2. – P. 314–334. – DOI: 10.1016/j.ymssp.2013.06.004
7. Babokin G.I., Shprekher D.M. Primenenie neyronnykh setey dlya diagnostiki elektromekhanicheskikh sistem // Gornyy inform.-analit. byul. (nauch.-tekhn. zhurn.). – 2011. – № S4. – S. 132–139.
8. Yunusova L.R., Magsumova A.R. Klasterizatsiya s pomoshch'yu neyronnykh setey i poisk zavisimostey // Nauka, obrazovanie i kul'tura. – 2019. – № 7(41). – S. 18–19.
9. Katser Yu.D., Kozitsin V.O., Maksimov I.V. Metody obnaruzheniya neispravnostey oborudovaniya AES // Izv. vuzov. Yadernaya energetika. – 2019. – № 4. – S. 5–27. – DOI: 10.26583/npe.2019.4.01
10. Rannyaya diagnostika i prognozirovanie nadezhnosti promyshlennogo oborudovaniya na osnove "tsifrovogo dvoynika" / S.A. Zhilyakov, E.M. Karasev, S.B. Levochkin, T.A. Pleshivtseva // Delovoy zhurn. Neftegaz.RU. – 2021. – № 5(113). – S. 60–66. – URL: https://magazine.neftegaz.ru/articles/tsifrovizatsiya/682121-rannyaya-diagnostika-i-prognozirovanie-nadezhnosti-promyshlennogo-oborudovaniya-na-osnove-tsifrovogo/ (data obrashcheniya 15.03.2023).
11. Korshikova A.A., Trofimov A.G. Model' rannego obnaruzheniya avariynykh situatsiy na oborudovanii elektrostantsiy na osnove metodov mashinnogo obucheniya // Teploenergetika. – 2019. – № 3. – S. 49–56. – URL: https://inctrl.ru/files/nodus_items/0000/0233/attaches/Model-for-Early-Detection_ru.pdf (data obrashcheniya 01.04.2023). – DOI: 10.1134/S0040363619030044
12. RUL Estimation in Rotary Machines Using Linear Dimension Reduction and Bayesian Inference / M. Heydarzadeh, M. Zafarani, B. Akin, M. Nourani // 2017 IEEE Int. Electric Machines and Drives Conf. (IEMDC), Miami, FL, USA, May 21–24, 2017. – IEEE, 2017. – DOI: 10.1109/IEMDC.2017.8002380
13. Shcherbatov I.A., Tsurikov G.N. Opredelenie defektov energeticheskogo oborudovaniya s pomoshch'yu algoritmov mashinnogo obucheniya v sistemakh prediktivnoy analitiki // Sovremennye problemy teplofiziki i energetiki: materialy III mezhdunar. konf., M., 19–23 okt. 2020 g. – M.: NIU "MEI", 2020. – S. 706–707.
14. Sarafanov M., Vychuzhanin P., Nikitin N. Prognozirovanie vremennykh ryadov s pomoshch'yu AutoML. – URL: https://habr.com/ru/post/559796/ (data obrashcheniya 10.09.2021).