Data quality analysis of the station of geological and technological researches in recognizing losses and kicks to improve the prediction accuracy of neural network algorithms

UDK: 681.518:622.24
DOI: 10.24887/0028-2448-2020-8-63-67
Key words: data quality, geological and technological studies, losses, kicks, neural network algorithms, prevention of accidents and troubles, artificial intelligence, automated system, well construction
Authors: A.I. Arkhipov (Oil and Gas Research Institute of RAS, RF, Moscow; Gubkin University, RF, Moscow), A.N. Dmitrievsky (Oil and Gas Research Institute of RAS, RF, Moscow; Gubkin University, RF, Moscow), N.A. Eremin (Oil and Gas Research Institute of RAS, RF, Moscow; Gubkin University, RF, Moscow), A.D. Chernikov (Oil and Gas Research Institute of RAS, RF, Moscow), S.O. Borozdin (Oil and Gas Research Institute of RAS, RF, Moscow; Gubkin University, RF, Moscow), E.A. Safarova (Oil and Gas Research Institute of RAS, RF, Moscow), M.R. Seinaroev (Oil and Gas Research Institute of RAS, RF, Moscow; Gubkin University, RF, Moscow)
The article summarizes and analyzes the quality of data received during the construction of wells. High quality and completeness of real-time drilling data have become key factors for improving the efficiency of data mining for decision-making. A combined architecture has been created that supports the latest computing technologies with high-frequency real-time data for creating intelligent alerts, as well as for remote monitoring of real-time data status for a large number of drilling rigs in the drilling control center. The quality of data is characterized by such metrics as completeness, accuracy, objectivity, timeliness of provision, source of origin, uniqueness, availability, format, and value. Of greatest interest for drilling are metrics such as completeness and accuracy. The classification of low-quality data is given. Examples of low-quality data from a geological and technological research station are considered. Criteria for the recognition of losses and kiks are formulated taking into account the quality of data for their further use in an automated system for preventing troubles and emergencies during the construction of oil and gas wells based on the use of artificial intelligence technologies and machine learning. When creating a high-performance automated system for preventing troubless and emergencies during the construction of oil and gas wells using artificial intelligence technology, the WITSML 2.0 data transfer protocol and the WITSML server are used. With a very large number of operations on the rig, transmitting up to 60,000 records in real time every second every day, it becomes necessary to use BigGeoData to predict drilling problems and discover hidden patterns. The use of artificial intelligence and machine learning models requires continuous improvement as drilling data changes. When using the WITSML big data transfer protocol, the task of monitoring the performance of artificial intelligence models becomes difficult due to the increase in the number of wells with real-time data, types of artificial intelligence models and types of data storage for drilling. The neural network methods described in this article for recognizing errors in the data of geological and technological measurement stations made it possible to achieve recognition of low-quality data in an automatic mode and increase the accuracy of forecasting complications.
References
1. Larionov A.S., Arkhipov A.I., Rodionov S.B., Well information is a growth point for the oil and gas business (In Russ.), Vestnik Assotsiatsii burovykh podryadchikov, 2015, no. 1, pp. 31–38.
2. Eremin N.A., Chernikov A.D., Sardanashvili O.N. et al., Digital technologies for well construction. Creation of a high-performance automated system for preventing complications and emergencies during the construction of oil and gas wells (In Russ.), Delovoy zhurnal Neftegaz.Ru, 2020, no. 4 (100), pp. 38–50.
3. Dmitrievskiy A.N., Duplyakin V.O., Eremin N.A., Kapranov V.V., Algorithm for creating a neural network model for classification in systems for preventing complications and emergencies in construction of oil and gas wells (In Russ.), Datchiki i sistemy, 2019, no. 12(243), pp. 3–10, DOI: 10.25728/datsys.2019.12.1
4. Ivlev A.P., Eremin N.A., Petrobotics: robotic drilling systems (In Russ.), Burenie i neft', 2018, no. 2, pp. 8–13.
5. Dmitrievsky A.D., Eremin N.A., Stolyarov E.V., Digital transformation of gas production, IOP Conference Series: Materials Science and Engineering (MSE), 2019, V. 700, DOI: 10.1088/1757-899x/700/1/012052.
6. Abukova L.A., Dmitrievskiy A.N., Eremin N.A., Digital modernization of Russian oil and gas complex (In Russ.), Neftyanoe khozyaystvo = Oil Industry, 2017, no. 10, pp. 54–58, DOI: 10.24887/0028-2448-2017-10-54-58.
7. Chen D.C.-K., Gaynor T., Comeaux B., Glass K., Hole quality: Gateway to efficient drilling, Proceedings of Offshore Technology Conference, 2002, January 1, DOI: 10.4043/14277-MS.
8. Svensson I., Wooley M., Halland T., Improving data quality in WITSML data, SPE-181038-MS, 2016, DOI:10.2118/181038-MS.
9. Nugraha B., Nair R., Muhammad K., Smart real time data transfer surveillance with edge computing and centralized remote monitoring system, Proceedings of International Petroleum Technology Conference, 2020, January 13, DOI: 10.2523/IPTC-19588-MS.
10. Mayani M.G., Baybolov T., Rommetveit R. et al., Optimizing drilling wells and increasing the operation efficiency using digital twin technology, SPE-199566-MS, 2020, DOI: 10.2118/199566-MS.
11. Alotaibi B., Aman B., Nefai M., Real-time drilling models monitoring using artificial intelligence, SPE-194807-MS, 2019, DOI: 10.2118/194807-MS.
12. Djamaluddin B., Prabhakar P., James B. et al., Real-time drilling operation activity analysis data modelling with multidimensional approach and column-oriented storage, SPE-194701-MS, 2019, DOI: 10.2118/194701-MS.
13. Singh K., Yalamarty S.S., Kamyab M., Cheatham C., Cloud-based ROP prediction and optimization in real time using supervised machine learning, Proceedings of Unconventional Resources Technology Conference, 2019, July 31, DOI: 10.15530/urtec-2019-343.
The article summarizes and analyzes the quality of data received during the construction of wells. High quality and completeness of real-time drilling data have become key factors for improving the efficiency of data mining for decision-making. A combined architecture has been created that supports the latest computing technologies with high-frequency real-time data for creating intelligent alerts, as well as for remote monitoring of real-time data status for a large number of drilling rigs in the drilling control center. The quality of data is characterized by such metrics as completeness, accuracy, objectivity, timeliness of provision, source of origin, uniqueness, availability, format, and value. Of greatest interest for drilling are metrics such as completeness and accuracy. The classification of low-quality data is given. Examples of low-quality data from a geological and technological research station are considered. Criteria for the recognition of losses and kiks are formulated taking into account the quality of data for their further use in an automated system for preventing troubles and emergencies during the construction of oil and gas wells based on the use of artificial intelligence technologies and machine learning. When creating a high-performance automated system for preventing troubless and emergencies during the construction of oil and gas wells using artificial intelligence technology, the WITSML 2.0 data transfer protocol and the WITSML server are used. With a very large number of operations on the rig, transmitting up to 60,000 records in real time every second every day, it becomes necessary to use BigGeoData to predict drilling problems and discover hidden patterns. The use of artificial intelligence and machine learning models requires continuous improvement as drilling data changes. When using the WITSML big data transfer protocol, the task of monitoring the performance of artificial intelligence models becomes difficult due to the increase in the number of wells with real-time data, types of artificial intelligence models and types of data storage for drilling. The neural network methods described in this article for recognizing errors in the data of geological and technological measurement stations made it possible to achieve recognition of low-quality data in an automatic mode and increase the accuracy of forecasting complications.
References
1. Larionov A.S., Arkhipov A.I., Rodionov S.B., Well information is a growth point for the oil and gas business (In Russ.), Vestnik Assotsiatsii burovykh podryadchikov, 2015, no. 1, pp. 31–38.
2. Eremin N.A., Chernikov A.D., Sardanashvili O.N. et al., Digital technologies for well construction. Creation of a high-performance automated system for preventing complications and emergencies during the construction of oil and gas wells (In Russ.), Delovoy zhurnal Neftegaz.Ru, 2020, no. 4 (100), pp. 38–50.
3. Dmitrievskiy A.N., Duplyakin V.O., Eremin N.A., Kapranov V.V., Algorithm for creating a neural network model for classification in systems for preventing complications and emergencies in construction of oil and gas wells (In Russ.), Datchiki i sistemy, 2019, no. 12(243), pp. 3–10, DOI: 10.25728/datsys.2019.12.1
4. Ivlev A.P., Eremin N.A., Petrobotics: robotic drilling systems (In Russ.), Burenie i neft', 2018, no. 2, pp. 8–13.
5. Dmitrievsky A.D., Eremin N.A., Stolyarov E.V., Digital transformation of gas production, IOP Conference Series: Materials Science and Engineering (MSE), 2019, V. 700, DOI: 10.1088/1757-899x/700/1/012052.
6. Abukova L.A., Dmitrievskiy A.N., Eremin N.A., Digital modernization of Russian oil and gas complex (In Russ.), Neftyanoe khozyaystvo = Oil Industry, 2017, no. 10, pp. 54–58, DOI: 10.24887/0028-2448-2017-10-54-58.
7. Chen D.C.-K., Gaynor T., Comeaux B., Glass K., Hole quality: Gateway to efficient drilling, Proceedings of Offshore Technology Conference, 2002, January 1, DOI: 10.4043/14277-MS.
8. Svensson I., Wooley M., Halland T., Improving data quality in WITSML data, SPE-181038-MS, 2016, DOI:10.2118/181038-MS.
9. Nugraha B., Nair R., Muhammad K., Smart real time data transfer surveillance with edge computing and centralized remote monitoring system, Proceedings of International Petroleum Technology Conference, 2020, January 13, DOI: 10.2523/IPTC-19588-MS.
10. Mayani M.G., Baybolov T., Rommetveit R. et al., Optimizing drilling wells and increasing the operation efficiency using digital twin technology, SPE-199566-MS, 2020, DOI: 10.2118/199566-MS.
11. Alotaibi B., Aman B., Nefai M., Real-time drilling models monitoring using artificial intelligence, SPE-194807-MS, 2019, DOI: 10.2118/194807-MS.
12. Djamaluddin B., Prabhakar P., James B. et al., Real-time drilling operation activity analysis data modelling with multidimensional approach and column-oriented storage, SPE-194701-MS, 2019, DOI: 10.2118/194701-MS.
13. Singh K., Yalamarty S.S., Kamyab M., Cheatham C., Cloud-based ROP prediction and optimization in real time using supervised machine learning, Proceedings of Unconventional Resources Technology Conference, 2019, July 31, DOI: 10.15530/urtec-2019-343.


Attention!
To buy the complete text of article (a format - PDF) or to read the material which is in open access only the authorized visitors of the website can. .

Mobile applications

Read our magazine on mobile devices

Загрузить в Google play

Press Releases

19.11.2020
09.11.2020
21.10.2020