Artificial Neural Network Modeling and Forecasting of Oil Reservoir Performance

  • Ehsan AmirianEmail author
  • Eugene Fedutenko
  • Chaodong Yang
  • Zhangxin Chen
  • Long Nghiem
Part of the Lecture Notes in Social Networks book series (LNSN)


A detailed uncertainty analysis on numerical flow simulation models preserving a robust and reliable model for an oil and gas reservoir is often deterministic, cumbersome, and expensive (manpower and time-consuming). Presence of a high-dimensional data space consisting of a large number of operational and geological parameters impedes practical decision-making and future performance prediction for oil and gas recovery processes. Thus, the rise of uncertainty-based reservoir development scenarios has provoked reservoir engineers to look for alternative modeling techniques that are capable of being reevaluated numerous times to examine the impact of specific variables or probing a range of scenarios on production profiles. This study reviews simulation data-driven proxy models that can be used instead of an actual reservoir simulator by using data interpolation in a high-dimensional parameter space.

The current study integrates a systematic data analysis and numerical flow simulations to create a comprehensive data set for different recovery processes, which entails different characteristics labeling reservoir heterogeneities and other pertinent operational constraints. This representative data set is then utilized as the building blocks for a cognitive data-driven proxy modeling workflow. Artificial and computational intelligence techniques are used to cognitively train a data-driven proxy model. The trained proxy will be subsequently employed to predict the production performance for the underlying process in presence of new reservoir development scenarios. The predictability of the designed cognitive proxy model is evaluated via comparing the results from proxy with those from a commercial simulator. Particularly, we compare the performance of two types of interpolation models: the radial basis function neural network (RBF NN) and the multilayer Levenberg-Marquardt neural network (ML LM NN). Results of studies of nine different cases of SAGD, black oil, and unconventional reservoirs will be presented to illustrate the current approach.

The presented results and performance characteristics associated with data-driven proxy models which can be reevaluated much faster than explicit models for the underlying process highlight the great potential of this modeling approach to be integrated directly into most existing reservoir management routines. This paper provides a viable tool to overcome challenges related to dynamic assessment of uncertainties during history matching of recovery processes and signifies the ability of cognitive proxy modeling in future performance prediction of oil and gas recovery processes.


SAGD Data-driven Proxy modeling Numerical flow simulation Artificial and computational intelligence 



Artificial and computational intelligence


Artificial neural network


Black oil


Global history matching error (%)


Hidden layer


History matching


Latin hypercube design






Neural network


Net present value


Objective function


Output layer


Radial basis function


Single layer


Steam-assisted gravity drainage


Uncertainty analysis



The authors would like to thank Computer Modeling Group Ltd. for permission to publish this paper. This research is supported by the NSERC/AIEES/Foundation CMG and AITF (iCORE) Chairs.


  1. 1.
    Kjaerulff, U. B., & Madsen, A. L. (2008). Bayesian networks and influence diagrams. Springer Science Business Media, 200, 114.Google Scholar
  2. 2.
    Mehrotra, K., Mohan, C. K., & Ranka, S. (1997). Elements of artificial neural networks. Cambridge, MA: MIT Press.Google Scholar
  3. 3.
    Poulton, M. M. (Ed.). (2001). Computational neural networks for geophysical data processing (Vol. 30). Amsterdam: Elsevier.Google Scholar
  4. 4.
    Suh, S. (2012). Practical applications of data mining. Sudbury, MA: Jones & Bartlett.Google Scholar
  5. 5.
    Chen, Z. (2002). Characteristic mixed discontinuous finite element methods for advection-dominated diffusion problems. Computers & Mathematics with Applications, 191, 2509–2538.Google Scholar
  6. 6.
    Chen, Z., Ewing, R. E., Kuznetsov, Y., Lazarov, R., & Maliassov, S. (1994). Multilevel preconditioners for mixed methods for second order elliptic problems. Numerical Linear Algebra and Applications, 3, 427–453.CrossRefGoogle Scholar
  7. 7.
    Chen, Z., Huan, G., & Ma, Y. (2006). Computational methods for multiphase flows in porous media. Computational science and engineering series (Vol. 2). Philadelphia, PA: SIAM.CrossRefGoogle Scholar
  8. 8.
    Purewal, S. (2005). Production forecasting in the oil and gas industry—Current methods and future trends. Business Briefing: Exploration & Production: The Oil & Gas Review (2), p. 12.Google Scholar
  9. 9.
    Bravo, C. E. et al. (2012). State-of-the-art application of artificial intelligence and trends in the E&P industry: A technology survey. In SPE Intelligent Energy International. Society of Petroleum Engineers.Google Scholar
  10. 10.
    Mohaghegh, S. (2000). Virtual-intelligence applications in petroleum engineering: Part I-artificial neural networks. Journal of Petroleum Technology, 52, 64–73.CrossRefGoogle Scholar
  11. 11.
    Saputelli, L., Malki, H., Canelon, J., & Nikolaou, M. (2002). A critical overview of artificial neural network applications in the context of continuous oil field optimization. In SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers.Google Scholar
  12. 12.
    Aminian, K., Ameri, S., Oyerokun, A., & Thomas, B. (2003). Prediction of flow units and permeability using artificial neural networks. In SPE Western Regional/AAPG Pacific Section Joint Meeting. Society of Petroleum Engineers.Google Scholar
  13. 13.
    Chaki, S., Verma, A. K., Routray, A., Mohanty, W. K., & Jenamani, M. (2014). Well tops guided prediction of reservoir properties using modular neural network concept: A case study from western onshore, India. Journal of Petroleum Science and Engineering, 123, 155–163.CrossRefGoogle Scholar
  14. 14.
    Tang, H., Meddaugh, W. S., & Toomey, N. (2011). Using an artificial-neural-network method to predict carbonate well log facies successfully. SPE Reservoir Evaluation & Engineering, 14, 35–44.CrossRefGoogle Scholar
  15. 15.
    Awoleke, O., & Lane, R. (2011). Analysis of data from the Barnett shale using conventional statistical and virtual intelligence techniques. SPE Reservoir Evaluation & Engineering, 14, 544–556.CrossRefGoogle Scholar
  16. 16.
    Lechner, J. P., Zangl, G. (2005) Treating uncertainties in reservoir performance prediction with neural networks. In SPE Europec/EAGE Annual Conference. Society of Petroleum Engineers.Google Scholar
  17. 17.
    Zangl, G., Graf, T., & Al-Kinani, A. (2006). A proxy modeling in production optimization. In SPE Europec/EAGE Annual Conference and Exhibition. Society of Petroleum Engineers.Google Scholar
  18. 18.
    Artun, E., Ertekin, T., Watson, R., & Miller, B. (2012). Designing cyclic pressure pulsing in naturally fractured reservoirs using an inverse looking recurrent neural network. Computers & Geosciences, 38, 68–79.CrossRefGoogle Scholar
  19. 19.
    Ayala, H. L. F., & Ertekin, T. (2007). Neuro-simulation analysis of pressure maintenance operations in gas condensate reservoirs. Journal of Petroleum Science & Engineering, 58, 207–226.CrossRefGoogle Scholar
  20. 20.
    Costa, L. A. N., Maschio, C., & Schiozer, D. J. (2014). Application of artificial neural networks in a history matching process. Journal of Petroleum Science and Engineering, 123, 30–45.CrossRefGoogle Scholar
  21. 21.
    Maschio, C., & Schiozer, D. J. (2014). Bayesian history matching using artificial neural network and Markov chain Monte Carlo. Journal of Petroleum Science and Engineering, 123, 62–71.CrossRefGoogle Scholar
  22. 22.
    Amirian, E., & Chen, Z. (2015). Practical application of data-driven modeling approach during waterflooding operations in heterogeneous reservoirs. Paper presented at the SPE Western Regional Meeting-Garden Grove. Society of Petroleum Engineers.Google Scholar
  23. 23.
    Amirian, E., Leung, J. Y., Zanon, S., & Dzurman, P. (2013) Data-driven modeling approach for recovery performance prediction in SAGD operations. In SPE Heavy Oil Conference-Canada, Calgary, AB, Canada. Society of Petroleum Engineers.Google Scholar
  24. 24.
    Amirian, E., Leung, J. Y., Zanon, S., & Dzurman, P. (2015). Integrated cluster analysis and artificial neural network modeling for steam-assisted gravity drainage performance prediction in heterogeneous reservoirs. Expert Systems with Applications, 42, 723–740.CrossRefGoogle Scholar
  25. 25.
    Fedutenko, E., Yang, C., Card, C., & Nghiem L. X. (2012) Forecasting SAGD process under geological uncertainties using data-driven proxy model. In SPE Heavy Oil Conference Canada. . Society of Petroleum Engineers.Google Scholar
  26. 26.
    Fedutenko, E., Yang, C., Card, C., & Nghiem, L. X. (2014). Time-dependent neural network based proxy modeling of SAGD process. In SPE Heavy Oil Conference-Canada. Society of Petroleum Engineers.Google Scholar
  27. 27.
    Popa, A. S., & Cassidy, S. D. (2012). Artificial intelligence for heavy oil assets: The evolution of solutions and organization capability. In: SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers.Google Scholar
  28. 28.
    Parada, C. H., & Ertekin, T. (2012). A new screening tool for improved oil recovery methods using artificial neural networks. In SPE Western Regional Meeting. Society of Petroleum Engineers.Google Scholar
  29. 29.
    Zerafat, M. M., Ayatollahi, S., Mehranbod, N., & Barzegari, D. (2011). Bayesian network analysis as a tool for efficient EOR screening. In SPE Enhanced Oil Recovery Conference. Society of Petroleum Engineers.Google Scholar
  30. 30.
    Holdaway, K. R. (2012). Oilfield data mining workflows for robust reservoir characterization: Part 2, SPE Intelligent Energy International. Long Beach, CA: Society of Petroleum Engineers.CrossRefGoogle Scholar
  31. 31.
    Mohammadpoor, M., Firouz, Q., Reza, A., & Torabi, F. (2012). Implementing simulation and artificial intelligence tools to optimize the performance of the CO2 sequestration in coalbed methane reservoirs. In Carbon Management Technology Conference, Orlando, FL.Google Scholar
  32. 32.
    McKay, M. D., Beckman, R. J., & Conover, W. J. (1979). Comparison of three methods for selecting values of input variables in the analysis of output from a computer code. Technometrics, 21, 239–245.Google Scholar
  33. 33.
    Myers, R. H., & Montgomery, D. C. (2002). Response surface methodology. Process and product optimization using designed experiments. New York, NY: Wiley.Google Scholar
  34. 34.
    Killough, J. E. (1995, January). Ninth SPE comparative solution project: A reexamination of black-oil simulation. In SPE Reservoir Simulation Symposium. Society of Petroleum Engineers.Google Scholar

Copyright information

© Springer Nature Switzerland AG 2018

Authors and Affiliations

  • Ehsan Amirian
    • 1
    Email author
  • Eugene Fedutenko
    • 2
  • Chaodong Yang
    • 2
  • Zhangxin Chen
    • 1
  • Long Nghiem
    • 2
  1. 1.Department of Chemical and Petroleum Engineering, Schulich School of EngineeringUniversity of CalgaryCalgaryCanada
  2. 2.Computer Modelling Group Ltd.CalgaryCanada

Personalised recommendations