Территория «НЕФТЕГАЗ». 2020; 1: 12-18
Новый подход к оценке вязкости сырой нефти на основе метода машинного обучения
Аннотация
В статье рассмотрены возможности прогностической оценки вязкости нефти с помощью искусственного интеллекта. Для анализа вязкости ненасыщенных, насыщенных и дегазированных нефтей различных месторождений были применены методы искусственной нейронной сети и опорных векторов. Набор исходных данных включал 300 лабораторных измерений образцов нефти, был получен с месторождений всего мира. 75 % экспериментальных данных были использованы для испытания предложенных моделей искусственной нейронной сети, в то время как оставшиеся 25 % были задействованы при тестировании модели на производительность (качество).
Результат исследования продемонстрировал превосходство моделей машинного обучения над существующими моделями оценки вязкости нефти по данным термодинамических исследований. Сравнительные результаты демонстрируют высокую точность моделей, созданных с помощью искусственной нейронной сети, по сравнению с другими методиками машинного обучения.
Кроме того, установлено, что предложенная методика прогностической оценки вязкости нефти позволяет осуществлять моделирование при наличии минимума исходных данных. При этом расчет производится с помощью простейших функций, что расширяет возможности применения методики.
В то же время отмечено, что если для расчета вязкости ненасыщенной нефти было собрано достаточное количество данных, то база данных для прогнозирования вязкости насыщенной и дегазированной нефти нуждается в расширении.
Список литературы
1. Ahmadloo F., Asghari K., Araghi M.M. Heavy Oil Viscosity Prediction Using Surface Response Methodology // Proceedings of the Canadian International Petroleum Conference. Petroleum Society of Canada, 2009.
2. Balabin R.M., Syunyaev R.Z. Petroleum Resins Adsorption onto Quartz Sand: Near Infrared (NIR) Spectroscopy Study // Journal of Colloid and Interface Science. 2007. Vol. 318. No. 2. P. 167–174.
3. Balabin R.M., Safieva R.Z., Lomakina-Rumyantseva E.I. Comparison of Linear and Nonlinear Calibration Models Based on Near Infrared (NIR) Spectroscopy Data for Gasoline Properties Prediction // Chemometrics and Intelligent Laboratory Systems. 2007. Vol. 88. No. 2. P. 183–188.
4. Balabin R.M., Syunyaev R.Z., Schmid T. et al. Asphaltene Adsorption onto an Iron Surface: Combined Near-Infrared (NIR), Raman, and AFM Study of the Kinetics, Thermodynamics, and Layer Structure // Energy Fuel. 2011. Vol. 25. No. 1. P. 189–196.
5. Ayoub M.A., Raja A.M., Al-Marhoun M.A. Evaluation of Below Bubble Point Correlations and Construction of a New Neural Network Model // Proceedings of the Asia Pacific Oil and Gas Conference and Exhibition. 2007. 10.2118/108439-MS.
6. Beal C. The Viscosity of Air, Water, Natural Gas, Crude Oil and Its Associated Gases at Oil Field Temperatures and Pressures // Transactions of the AIME. 1946. Vol. 165. No. 1. P. 94–115.
7. Beggs H.D., Robinson J.R. Estimating the Viscosity of Crude Oil Systems // Journal of Petroleum Technology. 1975. No. 27. P. 1140–1141.
8. Glaso O. Generalized Pressure-Volume-Temperature Correlations // Journal of Petroleum Technology. 1980. Vol. 32. No. 5. P. 785–795.
9. Kaye S. Offshore California Viscosity Correlations // COFRC. 1985. TS85000940.
10. Al-Khafaji A.H., Abdul-Majeed G.H., Hassoon S.F. Viscosity Correlation for Dead, Live and Undersaturated Crude Oils // Journal of Petroleum Research. 1987. No. 6. P. 1–16.
11. Ng J.T., Egbogah E.O. An Improved Temperature-Viscosity Correlation for Crude Oil Systems // Journal of Petroleum Science and Engineering. 1990. No. 5. P. 197–200.
12. Labedi R.M. PVT Correlations of the African Crudes. PhD thesis. USA: Colorado School of Mines; 1982.
13. Labedi R. Improved Correlations for Predicting the Viscosity of Light Crudes // Journal of Petroleum Science and Engineering. 1992. No. 8. P. 221–234.
14. Kartoatmodjo T., Jakarta P., Schmidt Z. Large Data Bank improves Crude Physical Property Correlations // Oil and Gas Journal (USA). 1994. No. 4. P. 51–55.
15. Petrosky Jr. G.E., Farshad F.F. Viscosity Correlations for Gulf of Mexico Crude Oils // Proceedings of SPE Production Operations Symposium. 1995.
16. Bennison T. Prediction of Heavy Oil Viscosity // Presented at the IBC Heavy Oil Field Development Conference. 1998.
17. Elsharkawy A.M., Alikhan A.A. Models for Predicting the Viscosity of Middle East Crude Oils // Fuel. 1999. Vol. 78. No. 8. P. 891–903.
18. Whitson C.H., Brul M.R. Phase Behavior. Richardson, Texas: Henry L. Doherty Memorial Fund of AIME, Society of Petroleum Engineers Inc.; 2000.
19. Bergman D.F., Sutton R.P. An Update to Viscosity Correlations for Gas-Saturated Crude Oils // SPE Annual Technical Conference and Exhibition. 2007.
20. Hossain M.S., Sarica C., Zhang H.-Q. et al. Assessment and Development of Heavy Oil Viscosity Correlations // SPE International Thermal Operations and Heavy Oil Symposium. 2005.
21. Naseri A., Nikazar M., Mousavi Dehghani S.A. A Correlation Approach for Prediction of Crude Oil Viscosities // Journal of Petroleum Science and Engineering. 2005. No. 47. P. 163–174.
22. Hemmati-Sarapardeh A., Khishvand M., Naseri A., Mohammadi A.H. Toward Reservoir Oil Viscosity Correlation // Chemical Engineering Science. 2013. No. 90. P. 53–68.
23. El-Hoshoudy A.N., Ali O.I., Dessouky S.M. New Correlations for Prediction of Viscosity and Density of Egyptian Oil Reservoirs // Fuel. 2013. No. 112. P. 277–282.
24. Dindoruk B., Christman P.G. PVT Properties and Viscosity Correlations for Gulf of Mexico Oils // SPE Reservoir Evaluation & Engineering. 2004;7(6):427–437.
25. Alomair O., Elsharkawy A.M., Alkandari H.A. A Viscosity Prediction for Kuwaiti Heavy Crudes at Elevated Temperatures // SPE Heavy Oil Conference and Exhibition. 2011. P. 1–18.
Territorija “NEFTEGAS” [Oil and Gas Territory]. 2020; 1: 12-18
A New Approach to Estimating Crude Oil Viscosity Based on Machine Learning Method
Abstract
The article discusses the possibilities of predictive assessment of oil viscosity using artificial intelligence. To analyze the viscosity of unsaturated, saturated and degassed oils from various fields, methods of an artificial neural network and support vectors were applied. The baseline data set included 300 laboratory measurements of oil samples, was obtained from fields around the world. 75 % of the experimental data were used to test the proposed artificial neural network models, while the remaining 25 % were used to test the model for performance (quality).
The result of the study demonstrated the superiority of machine learning models over existing models for assessing oil viscosity based on thermodynamic research data. Comparative results demonstrate the high accuracy of models created using an artificial neural network compared to other machine learning techniques.
In addition, it was found that the proposed method for predicting oil viscosity allows modeling with a minimum of initial data. In this case, the calculation is performed using the simplest functions, which expands the possibilities of applying the technique.
At the same time, it was noted that if enough data were collected to calculate the viscosity of unsaturated oil, then the database for predicting the viscosity of saturated and degassed oil needs to be expanded.
References
1. Ahmadloo F., Asghari K., Araghi M.M. Heavy Oil Viscosity Prediction Using Surface Response Methodology // Proceedings of the Canadian International Petroleum Conference. Petroleum Society of Canada, 2009.
2. Balabin R.M., Syunyaev R.Z. Petroleum Resins Adsorption onto Quartz Sand: Near Infrared (NIR) Spectroscopy Study // Journal of Colloid and Interface Science. 2007. Vol. 318. No. 2. P. 167–174.
3. Balabin R.M., Safieva R.Z., Lomakina-Rumyantseva E.I. Comparison of Linear and Nonlinear Calibration Models Based on Near Infrared (NIR) Spectroscopy Data for Gasoline Properties Prediction // Chemometrics and Intelligent Laboratory Systems. 2007. Vol. 88. No. 2. P. 183–188.
4. Balabin R.M., Syunyaev R.Z., Schmid T. et al. Asphaltene Adsorption onto an Iron Surface: Combined Near-Infrared (NIR), Raman, and AFM Study of the Kinetics, Thermodynamics, and Layer Structure // Energy Fuel. 2011. Vol. 25. No. 1. P. 189–196.
5. Ayoub M.A., Raja A.M., Al-Marhoun M.A. Evaluation of Below Bubble Point Correlations and Construction of a New Neural Network Model // Proceedings of the Asia Pacific Oil and Gas Conference and Exhibition. 2007. 10.2118/108439-MS.
6. Beal C. The Viscosity of Air, Water, Natural Gas, Crude Oil and Its Associated Gases at Oil Field Temperatures and Pressures // Transactions of the AIME. 1946. Vol. 165. No. 1. P. 94–115.
7. Beggs H.D., Robinson J.R. Estimating the Viscosity of Crude Oil Systems // Journal of Petroleum Technology. 1975. No. 27. P. 1140–1141.
8. Glaso O. Generalized Pressure-Volume-Temperature Correlations // Journal of Petroleum Technology. 1980. Vol. 32. No. 5. P. 785–795.
9. Kaye S. Offshore California Viscosity Correlations // COFRC. 1985. TS85000940.
10. Al-Khafaji A.H., Abdul-Majeed G.H., Hassoon S.F. Viscosity Correlation for Dead, Live and Undersaturated Crude Oils // Journal of Petroleum Research. 1987. No. 6. P. 1–16.
11. Ng J.T., Egbogah E.O. An Improved Temperature-Viscosity Correlation for Crude Oil Systems // Journal of Petroleum Science and Engineering. 1990. No. 5. P. 197–200.
12. Labedi R.M. PVT Correlations of the African Crudes. PhD thesis. USA: Colorado School of Mines; 1982.
13. Labedi R. Improved Correlations for Predicting the Viscosity of Light Crudes // Journal of Petroleum Science and Engineering. 1992. No. 8. P. 221–234.
14. Kartoatmodjo T., Jakarta P., Schmidt Z. Large Data Bank improves Crude Physical Property Correlations // Oil and Gas Journal (USA). 1994. No. 4. P. 51–55.
15. Petrosky Jr. G.E., Farshad F.F. Viscosity Correlations for Gulf of Mexico Crude Oils // Proceedings of SPE Production Operations Symposium. 1995.
16. Bennison T. Prediction of Heavy Oil Viscosity // Presented at the IBC Heavy Oil Field Development Conference. 1998.
17. Elsharkawy A.M., Alikhan A.A. Models for Predicting the Viscosity of Middle East Crude Oils // Fuel. 1999. Vol. 78. No. 8. P. 891–903.
18. Whitson C.H., Brul M.R. Phase Behavior. Richardson, Texas: Henry L. Doherty Memorial Fund of AIME, Society of Petroleum Engineers Inc.; 2000.
19. Bergman D.F., Sutton R.P. An Update to Viscosity Correlations for Gas-Saturated Crude Oils // SPE Annual Technical Conference and Exhibition. 2007.
20. Hossain M.S., Sarica C., Zhang H.-Q. et al. Assessment and Development of Heavy Oil Viscosity Correlations // SPE International Thermal Operations and Heavy Oil Symposium. 2005.
21. Naseri A., Nikazar M., Mousavi Dehghani S.A. A Correlation Approach for Prediction of Crude Oil Viscosities // Journal of Petroleum Science and Engineering. 2005. No. 47. P. 163–174.
22. Hemmati-Sarapardeh A., Khishvand M., Naseri A., Mohammadi A.H. Toward Reservoir Oil Viscosity Correlation // Chemical Engineering Science. 2013. No. 90. P. 53–68.
23. El-Hoshoudy A.N., Ali O.I., Dessouky S.M. New Correlations for Prediction of Viscosity and Density of Egyptian Oil Reservoirs // Fuel. 2013. No. 112. P. 277–282.
24. Dindoruk B., Christman P.G. PVT Properties and Viscosity Correlations for Gulf of Mexico Oils // SPE Reservoir Evaluation & Engineering. 2004;7(6):427–437.
25. Alomair O., Elsharkawy A.M., Alkandari H.A. A Viscosity Prediction for Kuwaiti Heavy Crudes at Elevated Temperatures // SPE Heavy Oil Conference and Exhibition. 2011. P. 1–18.
