Neural network pressure identifier of a pumping unit with asynchronous electric drive of a centrifugal pump

The article considers the development of a pressure identifier for a pumping unit with an asynchronous electric drive. The widespread use of such installations determines the relevance of this subject.

The purpose of the study is to develop a pressure identifier for a pumping unit as a part of an electrical complex based on neural networks. The current pressure value is calculated using the values of the stator current and voltage of the asynchronous electric motor. In the article under consideration, direct neural networks and nonlinear autoregressive networks with external input parameters are used to determine pressure, in contrast to existing methods for indirectly determining pressure.. Furthermore, to increase accuracy and reduce noise, the values of stator current modules of the electric motor and the value of the stator current voltage module are supplied to the network input.

As a result, the author obtains structures of neural network pressure identifiers with asynchronous electric drives and also determines the sizes of the hidden layers using integral criteria. The article provides comparative pressure waveforms showing the operability of the neuroidentifier.

1. Stepanov S. E., Vasenin A. B., Gruzdev V. V. [et al.]. Tsifrovoy dvoynik sistemy “plast–skvazhina–elektroprivodnoy tsentrobezhnyy nasos” neftegazovykh mestorozhdeniy [Digital twin of the system “Remat–well–electric drive centrifugal pump” oil and gas deposits]. Avtomatizatsiya i IT v Neftegazovoy Oblasti. 2024. No. 2 (56). P. 16–26. EDN: WQBTNP. (In Russ.).

2. Wolstrup A. F., Spangenberg J., Yamamoto A. [et al.]. Advances in 3D printed electromechanical sensors: Performance comparison, trends, and future directions. Additive Manufacturing. 2025. Vol. 106. P. 104799. DOI: 10.1016/j.addma.2025.104799.

3. Patent No. 2476983 C1 Russian Federation, IPC H 02 P 23/14, H 02 P 27/04. Sposob opredeleniya otsenki chastoty vrashcheniya asinkhronnogo dvigatelya [A method for determining the estimate of the rotational speed of an asynchronous motor] / Glazy- rin A. S., Tkachuk R. Yu., Glazyrina T. A. [et al.]. No. 2011135823/07. EDN: LZDPIE. (In Russ.).

4. Han Y., Zou J., Gong B. [et al.]. The use of model-based voltage and current analysis for torque oscillation detection and improved condition monitoring of centrifugal pumps. Mechanical Systems and Signal Processing. 2025. Vol. 222. P. 111781. DOI: 10.1016/j.ymssp.2024.111781.

5. Kuz’menko A. A. Arkhitektury neyronnykh setey [Neural network architectures]. Infokommunikacionnye Technologii. 2022. Vol. 20, no. 2. P. 121–132. DOI: 10.18469/ikt.2022.20.2.15. EDN: UGDBAM. (In Russ.).

6. Slavutskiy A. L., Slavutskiy L. A., Alekseyev V. V. [et al.]. Neyrosetevoy algoritm vosstanovleniya v real’nom vremeni signala promyshlennoy chastoty pri nelineynykh iskazheniyakh [A neural-network algorithm for real-time recovery of an industrialfrequency signal upon nonlinear distortions]. Elektrotekhnika. Electrotechnics. 2021. No. 8. P. 21–25. EDN: MDRKRB. (In Russ.).

7. Kuznetsov V. P. Kompleksnaya rekurrentnaya neyronnaya set’ [Complex recurrent neural network]. Vestnik Ryazanskogo gosudarstvennogo radiotekhnicheskogo universiteta. Vestnik of Ryazan State Radioengineering University. 2025. No. 92. P. 160–169. DOI: 10.21667/1995-4565-2025-92-160-169. EDN: HUBJMV. (In Russ.).

8. Leung H., Haykin S. The complex backpropagation algorithm. IEEE Transactions on Signal Processing. 1991. Vol. 39, no. 9. P. 2101–2104. DOI: 10.1109/78.134446.

9. Marcjasz G., Uniejewski B., Weron R. Probabilistic electricity price forecasting with NARX networks: Combine point or probabilistic forecasts. International Journal of Forecasting. 2020. Vol. 36, Issue 2. P. 466–479. DOI: 10.1016/j.ijforecast.2019.07.002.

10. Hagan M., Demuth H., Beale M. [et al.]. Neural network design. 2nd ed. Martin Hagan. 2014. P. 800. ISBN 978-0-97173211-7.

11. Jesús O. De., Hagan M. T. Backpropagation algorithms for a broad class of dynamic networks. IEEE Transactions on Neural Networks. 2007. Vol. 18, no. 1. P. 14–27. DOI: 10.1109/TNN.2006.882371. EDN: MEAGNJ.

12. Kochergin M. I. Realizatsiya neyronnykh setey v metode mnogourovnevykh komponentnykh tsepey [Implementation of neural networks in the method of multilevel component circuits]. Nauchno-tekhnicheskiy vestnik informatsionnykh tekhnologiy, mekhaniki i optiki. Scientific and Technical Journal of Information Technologies, Mechanics and Optics. 2023. Vol. 23, no. 6. P. 1162–1170. DOI: 10.17586/2226-1494-2023-23-6-1162-1170. EDN: SCZVOA. (In Russ.).