Abstract
We introduce a novel approach of data-driven dimensionality reduction for solving high-dimensional optimization problems, including history matching. Objective-Sensitive parameterization of the argument accounts for the corresponding change of objective function value. The result is achieved via an extension of the conventional loss function, which only quantifies approximation error over realizations. This paper contains three instances of such an approach based on Principal Component Analysis (PCA). Gradient-Sensitive PCA (GS-PCA) exploits a linear approximation of the objective function. Two other approaches solve the problem approximately within the framework of stationary perturbation theory (SPT). All the algorithms are verified and tested with a synthetic reservoir model. The results demonstrate improvements in parameterization quality regarding the reveal of the unconstrained objective function minimum. Also, we provide possible extensions and analyze the overall applicability of the Objective-Sensitive approach, which can be combined with modern parameterization techniques beyond PCA.
Similar content being viewed by others
Data Availability
The generated data is available from the corresponding author on reasonable request.
References
Aanonsen, S. I., Naevdal, G., Oliver, D. S., Reynolds, A. C., Valles, B.: The ensemble kalman filter in reservoir engineering-a review. Spe J. 14(3), 393–412 (2009). https://doi.org/10.2118/117274-pa
Bergstrm, D.: Surface generation & analysis – mysimlabs. http://www.mysimlabs.com/surface_generation.html(2012)
Brunton, S. L., Kutz, J. N.: Data-driven science and engineering: Machine learning, dynamical systems, and control. Cambridge University Press (2019)
Canchumuni, S.W., Emerick, A.A., Pacheco, M.A.C.: Towards a robust parameterization for conditioning facies models using deep variational autoencoders and ensemble smoother. Comput. Geosci. 128, 87–102. https://doi.org/10.1016/j.cageo.2019.04.006 (2019)
Carlberg, K., Bou-Mosleh, C., Farhat, C.: Efficient non-linear model reduction via a least-squares petrov-galerkin projection and compressive tensor approximations. Int. J. Numer. Methods Eng. 86(2), 155–181 (2011). https://doi.org/10.1002/nme.3050
Chen, Y., Oliver, D. S.: History matching of the norne full-field model with an iterative ensemble smoother. Spe Reserv. Eval. Eng. 17(2), 244–256 (2014). https://doi.org/10.2118/164902-pa
Evensen, G., Hove, J., Meisingset, H., Reiso, E., Seim, K. S., Espelid, Ø., et al.: Using the Enkf for Assisted History Matching of a North Sea Reservoir Model. In: SPE Reservoir Simulation Symposium. Society of Petroleum Engineers (2007)
Hajizadeh, Y., Christie, M., Demyanov, V.: Ant colony optimization for history matching and uncertainty quantification of reservoir models. J. Pet. Sci. Eng. 77(1), 78–92 (2011). https://doi.org/10.1016/j.petrol.2011.02.005
Hajizadeh, Y., Christie, M. A., Demyanov, V., et al.: History Matching with Differential Evolution Approach; a Look at New Search Strategies. In: SPE EUROPEC/EAGE Annual Conference and Exhibition. Society of Petroleum Engineers (2010)
Hajizadeh, Y., Christie, M. A., Demyanov, V., et al.: Towards Multiobjective History Matching: Faster Convergence and Uncertainty Quantification. In: SPE Reservoir Simulation Symposium. Society of Petroleum Engineers (2011)
Jafarpour, B., McLaughlin, D. B.: History matching with an ensemble kalman filter and discrete cosine parameterization. Comput. Geosci. 12(2), 227–244 (2008). https://doi.org/10.1007/s10596-008-9080-3
Jansen, J. D.: Adjoint-based optimization of multi-phase flow through porous media - a review. Comput. Fluids 46(1), 40–51 (2011). https://doi.org/10.1016/j.compfluid.2010.09.039
Jiang, A., Jafarpour, B.: History matching under uncertain geologic scenarios with variational autoencoders 2020(1), 1–14. https://doi.org/10.3997/2214-4609.202035194. https://www.earthdoc.org/content/papers/10.3997/2214-4609.202035194 (2020)
Kadyrova, A., Khlyupin, A.: Application of Adjoint-Based Optimal Control to Gas Reservoir with a Memory Effect. In: ECMOR XVI - 16Th European Conference on the Mathematics of Oil Recovery. https://doi.org/10.3997/2214-4609.201802214, vol. 2018, pp 1–13. European Association of Geoscientists & Engineers (2018)
Kaleta, M. P., Hanea, R. G., Heemink, A. W., Jansen, J. D.: Model-reduced gradient-based history matching. Comput. Geosci. 15(1), 135–153 (2011). https://doi.org/10.1007/s10596-010-9203-5
Khaninezhad, M.M., Jafarpour, B., Li, L.: Sparse geologic dictionaries for subsurface flow model calibration: Part i. inversion formulation. Adv. Water Resour. 39, 106–121. https://doi.org/10.1016/j.advwatres.2011.09.002. https://www.sciencedirect.com/science/article/pii/S0309170811001692 (2012)
Khaninezhad, M.M., Jafarpour, B., Li, L.: Sparse geologic dictionaries for subsurface flow model calibration: Part ii. robustness to uncertainty. Adv. Water Resour. 39, 122–136. https://doi.org/10.1016/j.advwatres.2011.10.005. https://www.sciencedirect.com/science/article/pii/S0309170811001977 (2012)
Khaninezhad, M.R.M., Jafarpour, B.: Field-scale history matching with sparse geologic dictionaries. J. Petrol. Sci. Eng. 170, 967–991. https://doi.org/10.1016/j.petrol.2018.06.024. https://www.sciencedirect.com/science/article/pii/S0920410518305126 (2018)
Laloy, E., Hérault, R., Lee, J., Jacques, D., Linde, N.: Inversion using a new low-dimensional representation of complex binary geological media based on a deep neural network. Adv. Water Resour. 110, 387–405. https://doi.org/10.1016/j.advwatres.2017.09.029. https://www.sciencedirect.com/science/article/pii/S0309170817306243 (2017)
Lie, K. A.: An introduction to reservoir simulation using MATLAB/GNU octave: User guide for the MATLAB reservoir simulation toolbox (MRST). Cambridge University Press. https://doi.org/10.1017/9781108591416 (2019)
Liu, Y. M., Sun, W. Y., Durlofsky, L. J.: A deep-learning-based geological parameterization for history matching complex models. Math. Geosci. 51(6), 725–766 (2019). https://doi.org/10.1007/s11004-019-09794-9
Lusch, B., Kutz, J.N., Brunton, S.L.: Deep learning for universal linear embeddings of nonlinear dynamics. Nat. Commun. 9, 10 (2018)
Mohamed, L., Christie, M. A., Demyanov, V., Robert, E., Kachuma, D., et al.: Application of Particle Swarms for History Matching in the Brugge Reservoir. In: SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers (2010)
Mukhin, A., Elizarev, M., Voskresenskiy, N., Khlyupin, A.: Application of dynamic parametrization algorithm for non-intrusive history matching approaches. https://doi.org/10.3997/2214-4609.202035045(2020)
Oliver, D. S., Chen, Y.: Recent progress on reservoir history matching: a review. Comput. Geosci. 15(1), 185–221 (2011). https://doi.org/10.1007/s10596-010-9194-2
Raissi, M., Perdikaris, P., Karniadakis, G.: Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys. 378, 686–707. https://doi.org/10.1016/j.jcp.2018.10.045. https://www.sciencedirect.com/science/article/pii/S0021999118307125 (2019)
Romero, C., Carter, J., Gringarten, A., Zimmerman, R., et al.: A Modified Genetic Algorithm for Reservoir Characterisation. In: International Oil and Gas Conference and Exhibition in China. Society of Petroleum Engineers (2000)
Sahni, I., Horne, R. N.: Multiresolution wavelet analysis for improved reservoir description. Spe Reserv. Eval. Eng. 8(1), 53–69 (2005). https://doi.org/10.2118/87820-pa
Sarma, P., Durlofsky, L. J., Aziz, K.: Kernel principal component analysis for efficient, differentiable parameterization of multipoint geostatistics. Math. Geosci. 40(1), 3–32 (2008). https://doi.org/10.1007/s11004-007-9131-7
Sarma, P., Durlofsky, L. J., Aziz, K., Chen, W. H.: Efficient real-time reservoir management using adjoint-based optimal control and model updating. Comput. Geosci. 10(1), 3–36 (2006). https://doi.org/10.1007/s10596-005-9009-z
Schmidt, E.: Zur Theorie Der Linearen Und Nichtlinearen Integralgleichungen. In: Integralgleichungen Und Gleichungen Mit Unendlich Vielen Unbekannten, pp 190–233. Springer (1989)
Schulze-Riegert, R. W., Axmann, J. K., Haase, O., Rian, D. T., You, Y. L.: Evolutionary algorithms applied to history matching of complex reservoirs. Spe Reserv. Eval. Eng. 5(2), 163–173 (2002). https://doi.org/10.2118/77301-pa
Skjervheim, J. A., Evensen, G., et al.: An Ensemble Smoother for Assisted History Matching. In: SPE Reservoir Simulation Symposium. Society of Petroleum Engineers (2011)
Van Essen, G., Jimenez, E., Przybysz-jarnut, J. K., Horesh, L., Douma, S. G., van den Hoek, P., Conn, A., Mello, U. T., et al.: Adjoint-Based History-Matching of Production and Time-Lapse Seismic Data. In: SPE Europec/EAGE Annual Conference. Society of Petroleum Engineers (2012)
Vo, H. X., Durlofsky, L. J.: A new differentiable parameterization based on principal component analysis for the low-dimensional representation of complex geological models. Math. Geosci. 46(7), 775–813 (2014). https://doi.org/10.1007/s11004-014-9541-2
Vo, H.X., Durlofsky, L.J.: Data assimilation and uncertainty assessment for complex geological models using a new pca-based parameterization. Comput. Geosci. 19(4), 747–767 (2015)
Vo, H. X., Durlofsky, L. J.: Regularized kernel pca for the efficient parameterization of complex geological models. J. Comput. Phys. 322, 859–881 (2016). https://doi.org/10.1016/j.jcp.2016.07.011
Wold, S., Esbensen, K., Geladi, P.: Principal component analysis. Chemometr. Intell. Labor. Syst. 2(1-3), 37–52 (1987)
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix: aGS-PCA evaluation
Perturbed eigenproblem evaluation
To obtain an approximate solution for GS-PCA, we consider the objective-sensitive optimal decomposition problem with a sensitivity parameter ε, which satisfies a small perturbation criterion.
Next, we express a residual norm \(||\boldsymbol {\mu }_{N \text {r}}||_{W}^{2}\) in terms of principal components φi and decomposition coefficients ai.
Assuming that the required solution is the first-order correction of standard principal components, we neglect higher-order terms of the scalar product \(\left <\boldsymbol {\varphi }_{i},\boldsymbol {\varphi }_{j} \right >_{W}\).
After that, we express the mean product < aiaj > in terms of principal components φi, second moment K, scalar product matrix W, and derive a corresponding expression for the mean residual norm \(<\|\boldsymbol {\mu }_{{N_{r}}}\|^{2}_{W}>\).
To achieve a sufficient simplicity of the approximate decomposition problem, we additionally neglect first-order perturbations in the obtained scalar product matrix WKW and the problem constraint.
Such constraint optimization implies the minimization of a corresponding Lagrangian.
Evaluating the necessary condition of constraint extremum,
we derive a perturbed eigenproblem associated with the approximate GS-PCA.
Perturbed eigenproblem solution
A first-order term \(\boldsymbol {\varphi }_{i}^{1}\) can be represented as a linear combination of unperturbed principal components given by a transform coefficients αij.
Given the known properties of unperturbed eigenvectors, we substitute decomposed vectors φ in the perturbed eigenproblem and project the expression onto an unperturbed vector \(\boldsymbol {\varphi }_{n}^{0}\).
Finally, the required expressions for the unknowns \({\sigma _{n}^{1}}\) and αkn are obtained from two cases of a relation between indices k and n.
In this work, we additionally perform Gram – Schmidt orthogonalization on resulting vectors in terms of a scalar product \(\left <\boldsymbol {\varphi }_{i},\boldsymbol {\varphi }_{j} \right >_{W}\).
Rights and permissions
About this article
Cite this article
Elizarev, M., Mukhin, A. & Khlyupin, A. Objective-sensitive principal component analysis for high-dimensional inverse problems. Comput Geosci 25, 2019–2031 (2021). https://doi.org/10.1007/s10596-021-10081-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10596-021-10081-y
Keywords
- Principal component analysis
- Dimensionality reduction
- Inverse problems
- Optimization
- History matching
- Reservoir simulation