Fault diagnosis of natural gas pumping unit based on machine learning

Autor: Kozlenko, Mykola, Kuz, Mykola, Zamikhovska, Olena, Zamikhovskyi, Leonid
Jazyk: angličtina
Rok vydání: 2022
Předmět:
ISSN: 1474-6670
DOI: 10.5281/zenodo.7409179
Popis: This paper presents a method for fault detection of natural gas pumping unit. It is a very difficult object for diagnosis. A lot of combinations of technical equipment, different operational conditions, and other factors require design and implementation of reliable diagnosis methods. Acoustic signal based fault diagnosis of natural gas pumping units is well known and widely used in a number of applications. Statistical modeling and frequency analysis are among the most popular. In this paper, we share our experience in the use of the classification model based on an artificial multilayered dense feed forward neural network and a deep learning approach for software-implemented diagnosis of a GTK-25-i type of pumping unit. The paper reports the overall accuracy of 0.98 and minimum F1-score of 0.8. This is competitive compared to the latest industry research findings.
M. Kozlenko, M. Kuz, O. Zamikhovska, and L. Zamikhovskyi, "Fault diagnosis of natural gas pumping unit based on machine learning," 6th International Scientific and Practical Conference on Applied Systems and Technologies in the Information Society (AISTIS), V. Pleskach, V. Zosimov, and M. Pyroh, Eds. Taras Shevchenko National University of Kyiv, Kyiv, Ukraine, Sept. 30, 2022, pp. 271-274, doi: 10.5281/zenodo.7409180
{"references":["Józef Korbicz, \"Neural Networks and their Application in Fault Detection and Diagnosis,\" IFAC Proceedings Volumes, vol. 30, no. 18, pp. 367-372, 1997, https://doi.org/10.1016/S1474-6670(17)42429-3","N. K. Verma, R. Gupta, R. K. Sevakula and A. Salour, \"Signal transforms for feature extraction from vibration signal for air compressor monitoring,\" TENCON 2014 - 2014 IEEE Region 10 Conference, Bangkok, Thailand, 2014, pp. 1-6, doi: 10.1109/TENCON.2014.7022275","L. Zamikhovskyі, O. Zamikhovska, and V. Pavlyk, \"Research of the characteristics of acoustic processes using wavelet transformation for detecting a diagnostic sign of the technical state of gas pumping units,\" TAPR, vol. 1, no. 2, pp. 32–36, Feb. 2021. https://doi.org/10.15587/2706-5448.2021.224432","O. Saubanov, A. Valeev, \"The Monitoring System of an Actual Technical Condition for Pumping Units with Frequency Analysis,\" Procedia Engineering, vol. 176, pp. 144-149, 2017 https://doi.org/10.1016/j.proeng.2017.02.282","M. Gorbiychuk, O. Zamikhovska, L. Zamikhovsky, and V. Pavlyk, \"Development of the method for estimating the technical condition of gas pumping units by their accelerating characteristic\", EEJET, vol. 3, no. 2, pp. 48–57, Jun. 2020. https://doi.org/10.15587/1729-4061.2020.206476","Y. G. Habashi, \"Application of neural networks to diagnostics of pumping-rod units,\" 2012 IV International Conference \"Problems of Cybernetics and Informatics\" (PCI), Baku, Azerbaijan, 2012, pp. 1-2, doi: 10.1109/ICPCI.2012.6486338","A. Mazumdar and A. S. Rawat, \"Learning and Recovery in the ReLU Model,\" 2019 57th Annual Allerton Conference on Communication, Control, and Computing (Allerton), Monticello, IL, USA, 2019, pp. 108-115, doi: 10.1109/ALLERTON.2019.8919900","M. Kozlenko, O. Zamikhovska, V. Tkachuk, and L. Zamikhovskyi, \"Deep learning based fault detection of natural gas pumping unit,\" 2021 IEEE 12th International Conference on Electronics and Information Technologies (ELIT), 2021, pp. 71-75, doi: 10.1109/ELIT53502.2021.9501066","M. Kozlenko, O. Zamikhovska, and L. Zamikhovskyi, \"Software implemented fault diagnosis of natural gas pumping unit based on feedforward neural network,\" Eastern-European Journal of Enterprise Technologies, vol. 2, no. 2(110), pp. 99-109, Apr. 2021, doi: 10.15587/1729-4061.2021.229859","М. Козленко, \"Дослідження ефективності використання частотного ресурсу при формуванні широкосмугових сигналів,\" Наукові вісті Галицька академія, no. 1, 2010, pp. 32-37, doi: 10.5281/zenodo.5652270"]}
Databáze: OpenAIRE
načítá se...