Abstract
Machine-learning-based proxy models are often used to replace many of the flow simulations required in optimizations of subsurface flow processes. Because the optimizer continuously shifts the solution toward promising regions of search space, the proxy must provide predictions outside its original training range. This requires retraining, as proxy models typically have limited extrapolation capabilities. In this work, detailed procedures for continuous proxy retraining are presented and evaluated. The basic proxy is an artificial neural network, and it is implemented in an iterative Latin hypercube sampling optimization procedure. At many optimization iterations, a fraction of the required function evaluations are performed using the proxy, and a fraction are via flow simulation. These simulation results are used to update the proxy, resulting in a dynamic strategy. The procedure is applied to two-dimensional and three-dimensional models of oil production via water injection, with wells operating under bottom-hole pressure control. Results demonstrate the advantages of the dynamic proxy strategy relative to a standard (fully simulation-based) approach. For the three-dimensional case, for example, with about the same number of numerical simulations, the proxy-based method gives a net present value (NPV) that exceeds that of the standard approach by 3.25%. In addition, the proxy strategy requires less than a third of the numerical simulations to achieve the same NPV as the standard procedure. The general strategy should be applicable with different optimizers for a range of optimization problems.
Similar content being viewed by others
References
Alpak FO, Jain V (2021) Support-vector regression accelerated well location optimization: algorithm, validation, and field testing. Comput Geosci 25:2033–2054
Bottazzi F, Della Rossa E (2017) A functional data analysis approach to surrogate modeling in reservoir and geomechanics uncertainty quantification. Math Geosci 49(4):517–540
Christie MA, Blunt M (2001) Tenth SPE comparative solution project: a comparison of upscaling techniques. SPE Reserv Eval Eng 4(04):308–317
Costa LAN, Maschio C, Schiozer DJ (2014) Application of artificial neural networks in a history matching process. J Petrol Sci Eng 123:30–45
de Brito DU, Durlofsky LJ (2021) Field development optimization using a sequence of surrogate treatments. Comput Geosci 25(1):35–65
Gaspar B, Teixeira AP, Soares CG (2014) Assessment of the efficiency of kriging surrogate models for structural reliability analysis. Probab Eng Mech 37:24–34
Gaspar B, Teixeira AP, Soares CG (2017) Adaptive surrogate model with active refinement combining kriging and a trust region method. Reliab Eng Syst Saf 165:277291
Goda T, Sato K (2013) Global optimization of injection well placement toward higher safety of CO\(_2\) geological storage. Energy Procedia 37:4583–4590
Goda T, Sato K (2014) History matching with iterative Latin hypercube samplings and parameterization of reservoir heterogeneity. J Petrol Sci Eng 114:61–73
Golzari A, Sefat MH, Jamshidi S (2015) Development of an adaptive surrogate model for production optimization. J Petrol Sci Eng 133:677–688
Guo Z, Reynolds AC (2018) Robust life-cycle production optimization with a support-vector-regression proxy. SPE J 23(06):2409–2427
Haykin S (2010) Neural networks and learning machines, 3rd edn. Pearson Education Inc., New Jersey
Hunyinbo S, Ma Z, Leung JY (2021) Incorporating phase behavior constraints in the multi-objective optimization of a warm vaporized solvent injection process. J Petrol Sci Eng 205:108949
Jaber AK, Al-Jawad SN, Alhuraishawy AK (2019) A review of proxy modeling applications in numerical reservoir simulation. Arab J Geosci 12(22):1–16
Jansen JD, Durlofsky LJ (2017) Use of reduced-order models in well control optimization. Optim Eng 18(1):105–132
Jin G, Mendoza K, Roy B, Buswell DG (2019) Machine learning-based fracture-hit detection algorithm using LFDAS signal. Lead Edge 38(7):520–524
Kennedy J, Eberhart R (1995) Particle swarm optimization. In: Proceedings of ICNN’95 International Conference on Neural Networks, Vol 4, pp 1942-1948
Kim YD, Durlofsky LJ (2021) A recurrent neural network-based proxy model for well-control optimization with nonlinear output constraints. SPE J 26(04):1837–1857
Kingma DP, Ba J (2014) Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980
Ma Z, Leung J (2020) Design of warm solvent injection processes for heterogeneous heavy oil reservoirs: a hybrid workflow of multi-objective optimization and proxy models. J Petrol Sci Eng 191:107186
Ma Z, Leung JY (2020) A knowledge-based heterogeneity characterization framework for 3D steam-assisted gravity drainage reservoirs. Knowl-Based Syst 192:105327
Ma Z, Volkov O, Durlofsky LJ (2022) Multigroup strategy for well control optimization. J Petrol Sci Eng 214:110448
McKay MD, Beckman RJ, Conover WJ (1979) Comparison of three methods for selecting values of input variables in the analysis of output from a computer code. Technometrics 21(2):239–245
Nait Amar M, Zeraibi N, Redouane K (2018) Optimization of WAG process using dynamic proxy, genetic algorithm and ant colony optimization. Arab J Sci Eng 43(11):6399–6412
Nasir Y, Yu W, Sepehrnoori K (2020) Hybrid derivative-free technique and effiective machine learning surrogate for nonlinear constrained well placement and production optimization. J Petrol Sci Eng 186:106726
Panjalizadeh H, Alizadeh N, Mashhadi H (2014) A workflow for risk analysis and optimization of steam ooding scenario using static and dynamic proxy models. J Petrol Sci Eng 121:78–86
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Duchesnay E (2011) Scikit-learn: machine learning in Python. J Mach Learn Res 12:2825–2830
Picheny V, Ginsbourger D, Roustant O, Haftka RT, Kim N-H (2010) Adaptive designs of experiments for accurate approximation of a target region. J Mech Des, 132(7)
Qiao L, Wang H, Lu S, Liu Y, He T (2022) Novel self-adaptive shale gas production proxy model and its practical application. ACS Omega 7(10):8294–8305
Singh M, Srinivasan S (2020) Development of proxy model for hydraulic fracturing and seismic wave propagation processes. Math Geosci 52(1):81–110
Song J, Yang Y, Wu J, Wu J, Sun X, Lin J (2018) Adaptive surrogate model based multiobjective optimization for coastal aquifer management. J Hydrol 561:98–111
Storn R, Price K (1997) Differential evolution - a simple and efficient heuristic for global optimization over continuous spaces. J Global Optim 11(4):341–359
Wang G, Carr TR (2012) Marcellus shale lithofacies prediction by multiclass neural network classification in the Appalachian Basin. Math Geosci 44(8):975–1004
Zhou Y (2012) Parallel general-purpose reservoir simulation with coupled reservoir models and multisegment wells (Unpublished doctoral dissertation). Stanford University, Department of Energy Resources Engineering
Zou A, Ye T, Volkov O, Durlofsky LJ (2022) Effective treatment of geometric constraints in derivative-free well placement optimization. J Petrol Sci Eng 215:110635
Zubarev DI (2009) Pros and cons of applying proxy-models as a substitute for full reservoir simulations. Paper 124815 presented at SPE Annual Technical Conference and Exhibition
Acknowledgements
Zhiwei Ma thanks the Natural Sciences and Engineering Research Council of Canada (NSERC) Postdoctoral Fellowships Program for financial support. We are also grateful to the Stanford Smart Fields Consortium for additional funding. The computational resources used in this study were provided by the Stanford Center for Computational Earth and Environmental Sciences.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Ma, Z., Kim, Y.D., Volkov, O. et al. Optimization of Subsurface Flow Operations Using a Dynamic Proxy Strategy. Math Geosci 54, 1261–1287 (2022). https://doi.org/10.1007/s11004-022-10020-2
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11004-022-10020-2