Transparent open-box learning network and artificial neural network predictions of bubble-point pressure compared

Abstract

The transparent open box (TOB) learning network algorithm offers an alternative approach to the lack of transparency provided by most machine-learning algorithms. It provides the exact calculations and relationships among the underlying input variables of the datasets to which it is applied. It also has the capability to achieve credible and auditable levels of prediction accuracy to complex, non-linear datasets, typical of those encountered in the oil and gas sector, highlighting the potential for underfitting and overfitting. The algorithm is applied here to predict bubble-point pressure from a published PVT dataset of 166 data records involving four easy-to-measure variables (reservoir temperature, gas-oil ratio, oil gravity, gas density relative to air) with uneven, and in parts, sparse data coverage. The TOB network demonstrates high-prediction accuracy for this complex system, although it predictions applied to the full dataset are outperformed by an artificial neural network (ANN). However, the performance of the TOB algorithm reveals the risk of overfitting in the sparse areas of the dataset and achieves a prediction performance that matches the ANN algorithm where the underlying data population is adequate. The high levels of transparency and its inhibitions to overfitting enable the TOB learning network to provide complementary information about the underlying dataset to that provided by traditional machine learning algorithms. This makes them suitable for application in parallel with neural-network algorithms, to overcome their black-box tendencies, and for benchmarking the prediction performance of other machine learning algorithms.

Keywords

Learning network transparency / Learning network performance compared / Prediction of oil bubble point pressure / Over fitting data sets for prediction / Auditing machine learning predictions / TOB complements ANN

References

[1] J. Schmidhuber, Deep learning in neural networks: an overview, Neural Network. 61 (2015) 85-117.

[2] M. Heinert, Artificial neural networks -how to open the black boxes? in: A. Reiterer, U. Egly (Application of Artificial Intelligence in Engineering Geodesy.Eds.), Proceedings of AIEG December, vols. 42-62, 2008 978-3-9501492-4-1Vienna, Austria.

[3] D.A. Wood, A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms, Advances in Geo-Energy Research 2 (2) (2018) 148-162.

[4] S. Elkatatny, Z. Tariq, M. Mahmoud, Real time prediction of drilling fluid rheological properties using Artificial Neural Networks visible mathematical model (white box), J. Petrol. Sci. Eng. 146 (2016) 1202-1210.

[5] D.L. Katz, Prediction of shrinkage of crude oils. API, Drill Prod. Pract. (1942) 137-147.

[6] M.B. Standing, A Pressure-volume-temperature correlation for mixtures of California oils and gases. API, Drill Prod. Pract. (1947) 275-287.

[7] O. Glaso, Generalized pressure-volume-temperature correlations, J. Petrol. Technol. 32 (5) (1980) 785-795.

[8] M.A. Al-Marhoun, New correlation for formation volume factor of oil and gas mixtures, J. Can. Pet. Technol. 31 (1992) 22.

[9] M. Karimnezhad, M. Heidarian, M. Kamari, H. Jalalifar, A new empirical correlation for estimating bubble point oil formation volume factor, J. Nat. Gas Sci. Eng. 18 (2014) 329-335.

[10] A. Jarrahian, J. Moghadasi, E. Heidaryan, Empirical estimating of black oils bubble point (saturation) pressure, J. Petrol. Sci. Eng. 126 (2015) 69-77.

[11] G. Kim, J. Park, M. Lee, Bubble point measurement and high-pressure distillation column design for the environmentally benign separation of zirconium from hafnium for nuclear power reactor, Kor. J. Chem. Eng. 32 (1) (2015) 30-36.

[12] R.B. Gharbi, A.M. Elsharkawy, Neural-network model for estimating the PVT properties of Middle East crude oils, In: Paper SPE 37695 Presented at SPE Middle East Oil Show and Conference, Bahrain 15-18 (1997) March.

[13] N. Varotsis, V. Gaganis, J. Nighswander, P. Guieze,A novel noniterative method for the prediction of the PVT behavior of reservoir fluids, Paper SPE 56745 Presented at SPE Annual Technical Conference and Exhibition, Houston, Texas, 3-6 October, 1999.

[14] A.M. Malallah, R. Gharbi, M. Algharaib, Accurate estimation of the world crude oil PVT properties using graphical alternating conditional expectation, Energy Fuel. 20 (2006) 688-698.

[15] S. Dutta, J.P. Gupta, PVT correlations for Indian crude oil using artificial neural networks, J. Petrol. Sci. Eng. 72 (2010) 93-109.

[16] M.A. Al-Marhoun, S.S. Ali, A. Abdulraheem, S. Nizamuddin, A. Muhammadain, Prediction of bubble point pressure from composition of black oils using artificial neural network, Petrol. Sci. Technol. 32 (14) (2014) 1720-1728.

[17] M.A. Ahmadi, M. Pournik, S.R. Shadizadeh, Toward connectionist model for predicting bubble point pressure of crude oils: application of artificial intelligence, Petroleum 1 (4) (2014) 307-317.

[18] M.M. Mahmood, M.A. Al-Marhoun, Evaluation of empirically derived PVT properties for Pakistani crude oils, J. Petrol. Sci. Eng. 16 (1996) 275-290.

[19] M.H. Rammay, A. Abdulraheem, PVT correlations for Pakistani crude oils using artificial neural network, J Petrol Explor Prod Technol 7 (2017) 217-233.

[20] C.M. Bishop,Neural Networks for Pattern Recognition, second ed.ed., Oxford University Press. U.K., 0198538642, 1995, p. 482.

[21] S. Haykin, New York, U.S.A., Neural Networks:a Comprehensive Introduction, third ed.ed., Pearson /Prentice Hall, 1999, p. 906 ISBN-10:0-13-147139-2.

[22] D.A. Wood, A. Choubineh, B. Vaferi, Transparent open-box learning network provides auditable predictions: pool boiling heat transfer coefficient for alumina-waterbased nanofluids, J. Therm. Anal. Calorim. (2018) Published online 12 September 2018 https://doi.org/10.1007/s10973-018-7722-9.

[23] M. Birattari, G. Bontempi, H. Bersini, Lazy Learning Meets the Recursive Least Squares Algorithm. Advances in Neural Information Processing Systems vol. 11, MIT Press, Cambridge, MA, 1999, pp. 375-381 1999.

[24] G.H. Chen, D. Shah, Explaining the success of nearest neighbor methods in prediction, Foundations and Trends R in Machine Learning 10 (5-6) (2018) 337-588.

[25] R. Samworth, Optimal weighted nearest neighbour classifiers, Ann. Stat. 40 (5) (2012) 2733-2763.

[26] Frontline Solvers. Standard Excel Solver -Limitations of Nonlinear Optimization (accessed December 2018) https://www.solver.com/standard-excel-solverlimitations-nonlinear-optimization.