Газовая Промышленность 7.2023

Научная статья

УДК 004.89::622.276
(UDK 004.89::622.276)

Для получения доступа к статьям

Авторизуйтесь

ЦИФРОВИЗАЦИЯ (DIGITALIZATION)

ПОДХОД К РАСПОЗНАВАНИЮ ПРЕДАВАРИЙНЫХ СИТУАЦИЙ ПРИ ЭКСПЛУАТАЦИИ СКВАЖИН С ИСПОЛЬЗОВАНИЕМ МАЛЫХ ОБУЧАЮЩИХ ВЫБОРОК

(APPROACH TO RECOGNITION OF PRE-EMERGENCY SITUATIONS DURING WELL OPERATION USING SMALL TRAINING SAMPLES)

Цифровая трансформация затрагивает все процессы в нефтегазовой отрасли. Значительный положительный эффект от внедрения методов искусственного интеллекта удается получить при эксплуатации скважин. В статье предложен метод распознавания предаварийных ситуаций при работе нефтегазовых скважин с использованием обучающих выборок, в семь раз меньших по объему, чем в опубликованных ранее источниках. В исследовании использовался открытый набор данных 3W с редкими нежелательными событиями на морских нефтяных скважинах.
Информативные признаки извлекались из временн х рядов телеметрии с помощью многослойной сверточной нейронной сети типа автокодировщик.
Описан способ предобработки временн х рядов, позволивший реализовать схему переноса обучения, при которой автокодировщик удалось обучить на речевых образах, взятых из открытых наборов данных TIMIT и VoxCeleb1. После обучения нейронная сеть позволяет извлекать информативные признаки из телеметрических образов. В качестве классификатора применялся наивный байесовский, а также его модифицированный вариант с использованием усовершенствованной формулы. В результате удалось достигнуть точности двухклассовой классификации на уровне 0,97 и многоклассовой – 0,91.

Digital transformation affects all processes in the oil and gas industry. A significant positive effect of artificial intelligence can be achieved at well operation. The paper proposes a method for recognition of pre-emergency situations in oil and gas wells’ operating using training datasets that are seven times smaller than those in previously published works. The study uses 3W open dataset of rare undesirable events which may happen at offshore oil wells. Informative features are extracted from telemetry time series using a multilayer convolutional neural network of the autoencoder type.
The paper describes a method of pre-processing time series to implement a learning transfer scheme where the autoencoder is trained using speech patterns taken from TIMIT and VoxCeleb1 open datasets. After training, the neural network allows for extraction of informative features from telemetry patterns. A naive Bayes classifier and its modified version employing an improved formula are used as classifiers. This results in a two-class classification accuracy of about 0.97 and a multiclass classification accuracy of about 0.91.

ПРЕДИКТИВНОЕ ОБСЛУЖИВАНИЕ, ИСКУССТВЕННЫЙ ИНТЕЛЛЕКТ, ИСКУССТВЕННАЯ НЕЙРОННАЯ СЕТЬ, ОБУЧЕНИЕ АВТОКОДИРОВЩИКА, АНАЛИЗ ТЕЛЕМЕТРИЧЕСКИХ ДАННЫХ, ФОРМУЛА БАЙЕСА, РАСПОЗНАВАНИЕ ОБРАЗОВ, СПЕКТРОГРАММА

PREDICTIVE MAINTENANCE, ARTIFICIAL INTELLIGENCE, ARTIFICIAL NEURAL NETWORK, AUTOENCODER TRAINING, TELEMETRY DATA ANALYSIS, BAYES’ THEOREM, PATTERN RECOGNITION, SPECTROGRAM

П.С. Ложников, д.т.н., доц., ООО «Газпром ВНИИГАЗ» (Санкт-Петербург, Россия), ФГАОУ ВО «Омский государственный технический университет» (Омск, Россия), lozhnikov@mail.ru

С.В. Жоров, ООО «Газпром трансгаз Сургут» (Сургут, Россия), zhorovsv@yandex.ru

С.А. Клиновенко, ФГАОУ ВО «Омский государственный технический университет», sergey.klinovenko@gmail.com

Л.В. Плетнев, ООО «Газпром ВНИИГАЗ», pletnevlv@gmail.com

А.Е. Сулавко, к.т.н., ФГАОУ ВО «Омский государственный технический университет», sulavich@mail.ru

З.Н. Шандрыголов, к.т.н., ООО «Газпром ВНИИГАЗ», z.shan@yandex.ru

P.S. Lozhnikov, DSc in Engineering, Associate Professor, Gazprom VNIIGAZ LLC (Saint Petersburg, Russia), Omsk State Technical University (Omsk, Russia), lozhnikov@mail.ru

S.V. Zhorov, Gazprom transgaz Surgut (Surgut, Russia), zhorovsv@yandex.ru

S.A. Klinovenko, Omsk State Technical University, sergey.klinovenko@gmail.com

L.V. Pletnev, Gazprom VNIIGAZ LLC, pletnevlv@gmail.com

A.E. Sulavko, PhD in Engineering, Omsk State Technical University, sulavich@mail.ru

Z.N. Shandrygolov, PhD in Engineering, Gazprom VNIIGAZ LLC, z.shan@yandex.ru

Yadav G., Paul K. Architecture and security of SCADA systems: A review // International Journal of Critical Infrastructure Protection. 2021. Vol. 34. Article ID 100433. DOI: 10.1016/j.ijcip.2021.100433.

Sircar A., Yadav K., Rayavarapu K., et al. Application of machine learning and artificial intelligence in oil and gas industry // Pet. Res. 2021. Vol. 6, No. 4. P. 379–391. DOI: 10.1016/j.ptlrs.2021.05.009.

Zhang D., Chen Y., Meng J. Synthetic well logs generation via Recurrent Neural Networks // Petroleum Exploration and Development. 2018. Vol. 45, No. 4. P. 629–639. DOI: 10.1016/S1876-3804(18)30068-5.

Zhou D., Huang D., Hao J., et al. Vibration-based fault diagnosis of the natural gas compressor using adaptive stochastic resonance realized by Generative Adversarial Networks // Eng. Failure Anal. 2020. Vol. 116. Article ID 104759. DOI: 10.1016/j.engfailanal.2020.104759.

Wong P., Wong W.K., Juwono F.H., et al. A minimalist approach for detecting sensor abnormality in oil and gas platforms // Pet. Res. 2022. Vol. 7, No. 2. P. 177–185. DOI: 10.1016/j.ptlrs.2021.09.007.

Alguliyev R., Imamverdiyev Y., Sukhostat L. Intelligent diagnosis of petroleum equipment faults using a deep hybrid model // SN Appl. Sci. 2020. Vol. 2. Article ID 924. DOI: 10.1007/s42452-020-2741-0.

Li Y., Ge T. Imminence monitoring of critical events: A representation learning approach // SIGMOD '21: Proceedings of the 2021 International Conference on Management of Data. New York, NY, USA: ACM, 2021. P. 1103–1115. DOI: 10.1145/3448016.3452804.

Vargas R.E.V., Munaro C.J., Ciarelli P.M., et al. A realistic and public dataset with rare undesirable real events in oil wells // J. Pet. Sci. Eng. 2019. Vol. 181. Article ID 106223. DOI: 10.1016/j.petrol.2019.106223.

Marins M.A., Barros B.D., Santos I.H., et al. Fault detection and classification in oil wells and production/service lines using random forest // J. Pet. Sci. Eng. 2021. Vol. 197. Article ID 107879. DOI: 10.1016/j.petrol.2020.107879.

Carvalho B.G. Evaluating machine learning techniques for detection of flow instability events in offshore oil wells: MSc thesis. Vitória, Brazil: Federal University of Espírito Santo, 2021. 58 p.

Klinovenko S.A., Lozhnikov P.S., Sulavko A.E., Pletnev L.V. Recognition of pre-emergency situations during the operation of oil wells based on convolutional neural networks and a modified Bayes classifier // Proceedings of the Conference of Russian Young Researchers in Electrical and Electronic Engineering. Saint Petersburg: Saint Petersburg Electrotechnical University “LETI”, 2023. P. 1297–1302.

Sulavko A.E., Samotuga A.E., Kuprik I.A. Personal identification based on acoustic characteristics of outer ear using cepstral analysis, Bayesian classifier, and artificial neural networks // IET Biometrics. 2021. Vol. 10, No. 6. P. 692–705. DOI: 10.1049/bme2.12037.

Lukic Y., Vogt C., Dürr O., Stadelmann T. Speaker identification and clustering using convolutional neural networks // Proceedings of the 26th International Workshop on Machine Learning for Signal Processing (MLSP). Salerno, Italy: IEEE, 2016. P. 1–6. DOI: 10.1109/MLSP.2016.7738816.

Nagrani A., Chung J.S., Xie W., Zisserman A. Voxceleb: Large-scale speaker verification in the wild // Computer Speech & Language. 2020. Vol. 60. Article ID 101027. DOI: 10.1016/j.csl.2019.101027.

Sulavko A.E., Samotuga A.E., Stadnikov D.G., et al. Biometric authentication on the basis of electroencephalograms parameters // J. Phys.: Conf. Ser. 2019. Vol. 1260, No. 2. Article ID 022011 DOI: 10.1088/1742-6596/1260/2/022011.

Yadav G, Paul K. Architecture and security of SCADA systems: A review. International Journal of Critical Infrastructure Protection. 2021; 34: article ID 100433. https://doi.org/10.1016/j.ijcip.2021.100433.

Sircar A, Yadav K, Rayavarapu K, Bist N, Oza H. Application of machine learning and artificial intelligence in oil and gas industry. Pet. Res. 2021; 6(4): 379–391. https://doi.org/10.1016/j.ptlrs.2021.05.009.

Zhang D, Chen Y, Meng J. Synthetic well logs generation via Recurrent Neural Networks. Petroleum Exploration and Development. 2018; 45(4): 629–639. https://doi.org/10.1016/S1876-3804(18)30068-5.

Zhou D, Huang D, Hao J, Ren Y, Jiang P, Jia X. Vibration-based fault diagnosis of the natural gas compressor using adaptive stochastic resonance realized by Generative Adversarial Networks. Eng. Failure Anal. 2020; 116: article ID 104759. https://doi.org/10.1016/j.engfailanal.2020.104759.

Wong P, Wong WK, Juwono FH, Gopal L, Yusoff MA. A minimalist approach for detecting sensor abnormality in oil and gas platforms. Pet. Res. 2022; 7(2): 177–185. https://doi.org/10.1016/j.ptlrs.2021.09.007.

Alguliyev R, Imamverdiyev Y, Sukhostat L. Intelligent diagnosis of petroleum equipment faults using a deep hybrid model. SN Appl. Sci. 2020; 2: article ID 924. https://doi.org/10.1007/s42452-020-2741-0.

Li Y, Ge T. Imminence monitoring of critical events: A representation learning approach. In: ACM SIGMOD SIGMOD '21: Proceedings of the 2021 International Conference on Management of Data, 20–25 June 2021, Virtual event China. New York, NY, USA: ACM; 2021. p. 1103–1115. https://doi.org/10.1145/3448016.3452804.

Vargas REV, Munaro CJ, Ciarelli PM, Medeiros AM, do Amaral BG, Barrionuevo DC, et al. A realistic and public dataset with rare undesirable real events in oil wells. J. Pet. Sci. Eng. 2019; 181: article ID 106223. https://doi.org/10.1016/j.petrol.2019.106223.

Marins MA, Barros BD, Santos IH, Barrionuevo DC, Vargas REV, Prego TDM, et al. Fault detection and classification in oil wells and production/service lines using random forest. J. Pet. Sci. Eng. 2021; 197: article ID 107879. https://doi.org/10.1016/j.petrol.2020.107879.

Carvalho BG. Evaluating machine learning techniques for detection of flow instability events in offshore oil wells. MSc thesis. Federal University of Espírito Santo; 2021.

Klinovenko SA, Lozhnikov PS, Sulavko AE, Pletnev LV. Recognition of pre-emergency situations during the operation of oil wells based on convolutional neural networks and a modified Bayes classifier. In: Saint Petersburg Electrotechnical University “LETI” Proceedings of the Conference of Russian Young Researchers in Electrical and Electronic Engineering, 23–27 January 2023, Saint Petersburg, Russia. Saint Petersburg: Saint Petersburg Electrotechnical University “LETI”; 2023. p. 1297–1302.

Sulavko AE, Samotuga AE, Kuprik IA. Personal identification based on acoustic characteristics of outer ear using cepstral analysis, Bayesian classifier, and artificial neural networks. IET Biometrics. 2021; 10(6): 692–705. https://doi.org/10.1049/bme2.12037.

Lukic Y, Vogt C, Dürr O, Stadelmann T. Speaker identification and clustering using convolutional neural networks. In: IEEE Proceedings of the 26th International Workshop on Machine Learning for Signal Processing (MLSP), 13–16 September 2016, Salerno, Italy. Salerno, Italy: IEEE; 2016. p. 1–6. https://doi.org/10.1109/MLSP.2016.7738816.

Nagrani A, Chung JS, Xie W, Zisserman A. Voxceleb: Large-scale speaker verification in the wild. Computer Speech & Language. 2020; 60: article ID 101027. https://doi.org/10.1016/j.csl.2019.101027.

Sulavko AE, Samotuga AE, Stadnikov DG, Pasenchuk VA, Zhumazhanova SS. Biometric authentication on the basis of electroencephalograms parameters. J. Phys.: Conf. Ser. 2019; 1260(2): article ID 022011 https://doi.org/10.1088/1742-6596/1260/2/022011.
NEFTEGAS.info

Внимание к деталям — от идеи
до воплощения! Только актуальная информация и свежие новости.

Контакты

108811, г. Москва, Киевское ш.,
Бизнес-парк «Румянцево», корп. Б,
подъезд 5, офис 506 Б

+7 (495) 240-54-57