Skip to main content
SpringerLink
Log in

Artificial neural networks workflow and its application in the petroleum industry

  • Review
  • Published:
Neural Computing and Applications Aims and scope Submit manuscript

Abstract

We develop a neural network workflow, which provides a systematic approach for tackling various problems in petroleum engineering. The workflow covers several design issues for constructing neural network models, especially in terms of developing the network structure. We apply the model to predict water saturation in an oilfield in Oman. Water saturation can be accurately obtained from data measured from cores removed from the oil field, but this information is limited to a few wells. Wireline log data are more abundantly available in most wells, and they provide valuable, but indirect, information about rock properties. A three-layered neural network model with five hidden neurons and a resilient back-propagation algorithm is found to be the best design for the saturation prediction. The input variables to the model are density, neutron, resistivity, and photo-electric wireline logs, and the model is trained using core water saturation. The model is able to predict the saturation directly from wireline logs with a correlation coefficient (r) of 0.91 and an error of 2.5 saturation units on the testing data.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12
Fig. 13
Fig. 14

Similar content being viewed by others

References

  1. Ali JK (1994) Neural network: a new tool for petroleum industry. Proceedings of SPE European petroleum computer conference, UK, paper 27561

  2. Olson TM (1998) Porosity and permeability prediction in low-permeability gas reservoirs from well logs using neural networks. Proceedings of SPE rocky mountain regional/low-permeability reservoir symposium and exhibition, USA, paper 39964

  3. White AC, Molnar D, Aminian K, Mohaghegh S, Ameri S (1995) The application of ANN for zone identification in a complex reservoir. Proceedings of the SPE Eastern regional conference & exhibition, USA, paper 30977

  4. Bhatt A, Helle HB (2002) Determination of facies from well logs using modular neural networks. Pet Geosci 8:217–228

    Article  Google Scholar 

  5. Helle HB, Bhatt A (2002) Fluid saturation from well logs using committee neural networks. Pet Geosci 8:109–118

    Article  Google Scholar 

  6. Al-Bulushi N, King P, Blunt M, Kraaijveld M (2009) Development of articial neural network for predicting water saturation and fluid distribution. J Petrol Sci Eng 68:197–208

    Article  Google Scholar 

  7. Archie GE (1941) The electrical resistivity log as an aid in determining some reservoir characteristics. Trans AIME 146:54–62

    Google Scholar 

  8. Waxman MH, Smits LJM (1968) Electrical conductivities in oil-bearing shaly sands. Soc Petrol Eng J 18(2):107–122

    Google Scholar 

  9. Leverett MC (1941) Capillary behavior in porous solids. Trans AIME 146:152–169

    Google Scholar 

  10. Worthington PE (1985) The evolution of shaly sand concepts in reservoir evaluation. The Log Analyst 26(1):23–40

    Google Scholar 

  11. Ipek I (2002) Log-derived cation exchange capacity of shaly sands: application to hydrocarbon detection and drilling optimisation. Doctoral diss., Louisiana State University

  12. Rezaee MR, Lemon NM (1996) Petrophysical evaluation of kaolinite-bearing sandstone: Water saturation (Sw), an example of the Tirrawarra sandstone reservoir, Copper Basin. Proceedings of the SPE Asia pacific oil & gas conference, Australia, paper 37023

  13. Haykin S (1998) neural networks, a comprehensive foundation. Prentice Hall PTR, 2nd edn. USA

  14. Essenreiter R, Karrenbach M, Treite S (1998) Multiple reflection attenuation in seismic data using back-propagation. IEEE Trans Signal Process 46(7):2001–2011

    Article  Google Scholar 

  15. Maren AJ, Harston CT, Pap RM (1990) Handbook of neural computing applications. Academic Press, New York

  16. Hush DR, Horne BG (1993) Progress in supervised neural networks. IEEE Signal Process Mag 10(1):8–39

    Article  Google Scholar 

  17. Bishop CM (1995) Neural networks for pattern recognition. Oxford University Press

  18. Hagan M, Demuth HB, Beale M (1996) Neural network design. PWS Publishing Company, USA

  19. Reidmiller M, Braun H (1993) A Direct adaptive method for faster back-propagation learning: the PROP algorithm. Proceedings of the IEEE international conference on neural network, pp 586–591

  20. Cun YLe, Denker JS, Solla SA (1990) Optimal brain damage. Adv Neural Network Process Syst 2:168–177

    Google Scholar 

  21. Hush DR, Horne B, Salas JM (1992) Error surfaces for multilayer perceptrons. IEEE Trans Syst Man Cybern 22 (5)

  22. Baum EB, Haussler D (1990) What size net gives valid generalization? Neural Comput 1(1):151–160

    Article  Google Scholar 

  23. Hush DR (1989) Classification with neural networks: a performance analysis. IEEE Int Conf Syst Eng, pp 227–280

  24. Hornik K, Stinchcombe M, White H (1989) Multilayer feedforward networks are universal approximates. Neural Netw 2:359–366

    Article  Google Scholar 

  25. Irie B, Miyake S (1998) Capabilities of there-layered perceptron. Proc IEEE Int Conf Neural Netw, pp 641–648

  26. Hecht-Nielsen R (1990) Neurocomputing. Addison-Wesley, New York

    Google Scholar 

  27. Huang SC, Huang YF (1991) Bounds on the number of hidden neurons in multilayer perceptrons. IEEE Trans Neural Netw 2(1):47–55

    Article  Google Scholar 

  28. Cheng B, Titterington DM (1994) Neural networks: a review from a statistical perspective. Stat Sci 9(1):2–54

    Article  MathSciNet  MATH  Google Scholar 

  29. Chester DL (1990) Why two hidden layers are better than one. Neural Netw J 1:265–268

    Google Scholar 

  30. Flood I, Kartam N (1994) Neural networks in civil engineering. I: Principles and understanding. J Comput Civil Eng 8(2):131–148

    Article  Google Scholar 

  31. Maier HR, Dandy GC (2000) Neural Networks for the prediction and forecasting of water resources variables: a review of modeling issues and applications. J Environ Modeling Softw 15:101–124

    Article  Google Scholar 

  32. Moody JE (1992) The effective number of parameters: an analysis of generalization and regularization in non-linear learning system. Adv Neural Netw Process Syst 4:841–854

    Google Scholar 

  33. Rogers LL, Dowla FU (1994) Optimization of groundwater remediation using artificial neural networks with parallel solute transport modeling. Water Resour Res 30(2):457–481

    Article  Google Scholar 

  34. Bann VDM, Jutten C (2000) Neural networks in geophysical applications. Geophysics 65(4):1032–1047

    Article  Google Scholar 

  35. Masters T (1993) Practical neural network recipes in C ≶≶. Academic Press

  36. Amari S, Murata N, Muller KR, Finke M, Yang H (1997) Asymptotic statistical theory of overtraining and cross-validation. IEEE Trans Neural Netw 8(5):985–996

    Article  Google Scholar 

  37. Hanson SJ, Pratt L (1989) Comparing biases for minimal network construction with back-propagation. Advances in neural information processing systems 1, (Morgan Kaufmann Publishers Inc, pp 177–185)

  38. Lachtermacher G, Fuller JD (1994) Back-propagation in hydrological time series forecasting, stochastic and statistical methodology in hydrology and environmental engineering. In Hipel KW, McLeod AI, Panu US, Singh VP (eds) Kluwer Academic Publisher, Dordrecht, pp 229–242

  39. Goda HM, Maier HR, Behrenbruch P (2005) The development of an optimal artificial neural network model for estimating initial, irreducible water saturation- Australian reservoirs. Proceedings of the SPE Asia Pacific Oil & Gas conference and Exhibition, Indonesia, paper 93307

  40. Stunder M, Al-Tuwaini JS (2001) How data-driven modeling like neural networks can help to integrate different types of data into reservoir management. Proceedings of the society of petroleum engineers middle east oil show, paper 68163

  41. Flood I, Kartam N (1994) Neural networks in civil engineering. I: Principles and understanding. J Comput Civil Eng 8(2):131–148

    Article  Google Scholar 

  42. Burden FR, Brereton RG, Walsh PT (1997) Cross-validatory selection of test and validation sets in multivariate calibration and neural networks as applied to spectroscopy. Analyst 122(10):1015–1022

    Article  Google Scholar 

  43. Kaastra I, Boyd MS (1995) Forecasting futures trading volume using neural networks. J Futures Mark 15(8):953–970

    Article  Google Scholar 

  44. Tukey JW (1977) Exploratory data analysis. Addison-Wesley, New York

    MATH  Google Scholar 

  45. Burke LI, Ignizio JP (1992) Neural networks and operations research: an overview. Comput Oper Res 19:179–189

    Article  MATH  Google Scholar 

  46. Faraway J, Chatfield C (1998) Time series forecasting with neural networks: a comparative study using the airline data. Appl Stat 14(2):109–250

    Google Scholar 

  47. Fortin V, Ouarda TBMJ, Bobee B (1997) Comment on ‘‘the use of artificial neural networks for the prediction of water quality parameters’’ by Maier HR and Dandy GC. Water Resour Res 33(10):2423–2424

    Article  Google Scholar 

  48. Shi JJ (2000) Reducing prediction error by transforming input data for neural networks. J Comput Civil Eng 14(2):109–166

    Article  Google Scholar 

  49. Dimopoulos I, Chronopoulos J, Chronopoulou-Sereli A, Lek Sovan (1999) Neural network models to study lead concentration in grasses and permanent urban descriptors in Athens city (Greece). Ecol Modell 120:157–165

    Article  Google Scholar 

  50. Gevery M, Dimopoulos I, Lek Sovan (2003) Review and comparison of methods to study the contribution of variables in artificial neural network models. Ecol Modell 160:249–264

    Article  Google Scholar 

  51. Garson D (1991) Interpreting neural network connection weights. AL Expert l6(4): 46

  52. Lek S, Delacoste M, Baran P, Dimopulos I, Lauga J, Auglagnier S (1996) Application of neural network to modelling nonlinear relationships in ecology. Ecol Modell 90:9–52

    Article  Google Scholar 

  53. Schlumberger (1991) Log interpretation principles/application. Schlumberger educational series seventh printing

  54. Schlumberger oil glossary http://www.glossary.oilfield.slb.com

Download references

Author information

Authors and Affiliations

  1. Petroleum Development of Oman (PDO), P.O. Box 761, 113, Muscat, Oman

    N. I. Al-Bulushi

  2. Department of Earth Science and Engineering, Imperial College London, London, SW7 2AZ, UK

    P. R. King & M. J. Blunt

  3. Shell International, Rijswijk, Netherlands

    M. Kraaijveld

Authors
  1. N. I. Al-Bulushi
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. P. R. King
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. M. J. Blunt
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. M. Kraaijveld
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to N. I. Al-Bulushi.

Appendix [52, 53]

Appendix [52, 53]

Wireline logs are continuous measurements of downhole formation through electrical instruments. The logging is performed during and after drilling a well where the tool is lowered to the formation through electrical cables. The interpretation of wireline logs provide indirect valuable information of the formation that has been drilled, such as lithology, porosity, saturation, and permeability. There are many tools used for wireline logging, such as gamma ray detector, density and neutron, resistivity measurement, and sonic travel time.

Wireline logs are continuous electrical measurements of downhole formation through electrical instruments. The measurements are performed at each depth of required formation, typically at every 150 cm provided through a long band of paper. The logging is performed during and after drilling a well where the tool is lowered to the formation through electrical cables. The interpretation of wireline logs provide indirect valuable information of the formation which has been drilled, such as lithology, porosity, saturation, and permeability. There are many tools used for wireline logging, such as gamma ray detector, density and neutron, resistivity measurement, and sonic travel time.

The gamma ray tool is a passive logging tool that detects the natural radiation of gamma rays from the formation, which are result of high-energy electromagnetic radiation [53, 54]. The gamma ray tool gives information about the lithology of the formation where high gamma rays are related to shaly environment, whereas low readings are interpreted as clean sands. The density log tool emits gamma ray into the formation that collides with the electrons in the formation [53, 54]. In the process, the gamma rays are attenuated. The counts rates of the scatted gamma ray at a fixed distance from the source are inversely related to the electron density of the formation; consequently, the bulk density of the formation can be calculated. The photo-electric absorption index gives information about the lithology of the formation where the photo-electric measurement primarily response to the rock matrix [53, 54]. The neutron logging tool bombards the formation with high-energy neutron. The high-energy neutron interact with the nucleus of the atoms in the formation where each interaction causes lose of neutron energy [53, 54]. The hydrogen atoms has the same mass of the neutron, hence lowers the speed of the neutron significantly. The slowing down rate is determined by the hydrogen index of all components of the formation and formation fluids that contact a significant fraction of hydrogen. The distance over which the neutrons have traveled before they reach a lower-energy level is related to the amount of the hydrogen atoms present in the formation. A combination of density and neutron tool gives indication of the lithology of the formation besides the gamma ray. The resistivity logging tool basically measures the resistivity of the formation. By measuring the resistivity, the water saturation can be calculated.

The coring and core analysis provide direct measurements of petrophysical properties in the laboratory. The core basically represents a whole section of rock extracted from the drilled formation. In the laboratory, samples from the core are taken for physical measurements such as porosity, saturation, and permeability. The physical properties from the core analysis represent the ground truth, and they are compared to the wireline logs calculated petrophysical properties. However, the core is an expensive method and limited only to few wells in the formation, whereas the wireline logs are abundantly available in most of the wells in the formation. Water saturation can be determined directly from cores taken from a well in the field. The widely used laboratory method to determine the water saturation is the Dean-Stark method. In this method, the fluid saturation is determined by distillation of the water fraction and extraction of the oil fraction from a sample. The process starts by vaporizing the water in the sample by boiling the solvent. This vaporized water is then condensed and collected in a calibrated trap. At this stage, the volume of the water in the sample can be determined. The solvent is also condensed and then flows back over the sample to extract the oil. After the water and oil have been removed, the sample is dried. The weight of oil is calculated by the difference between the total loss in the sample weight and water weight removed from it. The loss in the sample weight is calculated by measuring the weight of the sample before and after extraction.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Al-Bulushi, N.I., King, P.R., Blunt, M.J. et al. Artificial neural networks workflow and its application in the petroleum industry. Neural Comput & Applic 21, 409–421 (2012). https://doi.org/10.1007/s00521-010-0501-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00521-010-0501-6

Keywords

Search

Navigation