Bayesian Optimization Algorithm Applied to Uncertainty Quantification
- Asaad Abdollahzadeh (Heriot-Watt University) | Alan Reynolds (Heriot-Watt University) | Michael Christie (Herit-Watt University) | David W. Corne (Heriot-Watt University) | Brian J. Davies (BP) | Glyn J.J. Williams (BP)
- Document ID
- Society of Petroleum Engineers
- SPE Journal
- Publication Date
- September 2012
- Document Type
- Journal Paper
- 865 - 873
- 2012. Society of Petroleum Engineers
- 1.7.5 Well Control, 5.5.8 History Matching, 7.2.3 Decision-making Processes, 5.3.2 Multiphase Flow
- 11 in the last 30 days
- 663 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 10.00|
|SPE Non-Member Price:||USD 30.00|
Prudent decision making in subsurface assets requires reservoir uncertainty quantification. In a typical uncertainty-quantification study, reservoir models must be updated using the observed response from the reservoir by a process known as history matching. This involves solving an inverse problem, finding reservoir models that produce, under simulation, a similar response to that of the real reservoir. However, this requires multiple expensive multiphase-flow simulations. Thus, uncertainty-quantification studies employ optimization techniques to find acceptable models to be used in prediction. Different optimization algorithms and search strategies are presented in the literature, but they are generally unsatisfactory because of slow convergence to the optimal regions of the global search space, and, more importantly, failure in finding multiple acceptable reservoir models. In this context, a new approach is offered by estimation-of-distribution algorithms (EDAs). EDAs are population-based algorithms that use models to estimate the probability distribution of promising solutions and then generate new candidate solutions.
This paper explores the application of EDAs, including univariate and multivariate models. We discuss two histogram-based univariate models and one multivariate model, the Bayesian optimization algorithm (BOA), which employs Bayesian networks for modeling. By considering possible interactions between variables and exploiting explicitly stored knowledge of such interactions, EDAs can accelerate the search process while preserving search diversity. Unlike most existing approaches applied to uncertainty quantification, the Bayesian network allows the BOA to build solutions using flexible rules learned from the models obtained, rather than fixed rules, leading to better solutions and improved convergence. The BOA is naturally suited to finding good solutions in complex high-dimensional spaces, such as those typical in reservoir-uncertainty quantification.
We demonstrate the effectiveness of EDA by applying the well-known synthetic PUNQ-S3 case with multiple wells. This allows us to verify the methodology in a well-controlled case. Results show better estimation of uncertainty when compared with some other traditional population-based algorithms.
|File Size||3 MB||Number of Pages||9|
Abdollahzadeh, A., Reynolds, A., Christie, M.A., Corne, D., Williams, G.,and Davies, B.J. 2011. Estimation of Distribution Algorithms Applied to HistoryMatching. Paper SPE 141161 presented at the SPE Reservoir Simulation Symposium,The Woodlands, Texas, USA, 21-23 February. http://dx.doi.org/10.2118/141161-MS.
Baluja, S. 1994. Population-Based Incremental Learning: A Method forIntegrating Genetic Search Based Function Optimization and CompetitiveLearning. Technical Report 865123, CMU-CS-94-163, Carnegie Mellon University,Pittsburgh, Pennsylvania.
Barker, J.W., Cuypers, M., and Holden, L. 2001. Quantifying Uncertainty inProduction Forecasts: Another Look at the PUNQ-S3 Problem. SPE J.6 (4): 433-441. SPE-74707-PA. http://dx.doi.org/10.2118/74707-PA.
Bos, C.F.M. 1999. Production Forecasting with Uncertainty Quantification:PUNQ-2. Technical Report NITG 99-255-A, Netherlands Institute of AppliedGeoscience (TNO), The Hague, The Netherlands.
Coats, K.H., Dempsey, J.R., and Henderson, J.H. 1970. A New Technique forDetermining Reservoir Description from Field Performance Data. SPE J.10 (1): 66-74. SPE-2344-PA. http://dx.doi.org/10.2118/2344-PA.
Demyanov, V., Subbey, S., and Christie, M. 2004. Neighbourhood Algorithmwith Geostatistical Simulations for Uncertainty Quantification ReservoirModeling: PUNQ-S3 Case study. Paper presented at the 9th European Conference onthe Mathematics of Oil Recovery, Cannes, France, 30 August-2 September.
Erbas, D. and Christie, M.A. 2007. Effect of Sampling Strategies onPrediction Uncertainty Estimation. Paper SPE 106229 presented at the SPEReservoir Simulation Symposium, Houston, 26-28 February. http://dx.doi.org/10.2118/106229-MS.
Etxeberria, R. and Larrañaga, P. 1999. Global optimization using Bayesiannetworks. Proc., Second Symposium on Artificial Intelligence (CIMAF-99),Havana, Cuba, March, 151-173.
Evensen, G., Hove, J., Meisingset, H.C., Reiso, E., Seim, K.S., and Espelid,S.O. 2007. Using the EnKF for Assisted History Matching of a North SeaReservoir Model. Paper SPE 106184 presented at the SPE Reservoir SimulationSymposium, Houston, 26-28 February. http://dx.doi.org/10.2118/106184-MS.
Floris, F.J.T., Bush, M.D., Cuypers, M., Roggero, F., and Syversveen, A.-R.2001. Methods for quantifying the uncertainty of production forecasts: Acomparative study. Pet. Geosci. 7 (Supplement, 1 May): 87-96.
Hajizadeh, Y. 2010. Ants Can Do History Matching. Paper SPE 141137 presentedat the SPE Annual Technical Conference and Exhibition, Florence, Italy, 19-22September. http://dx.doi.org/10.2118/141137-STU.
Hajizadeh, Y., Christie, M.A., and Demyanov, V. 2010. Comparative Study ofNovel Population-Based Optimization Algorithms for History Matching andUncertainty Quantification: PUNQ-S3 Revisited. Paper SPE 136861 presented atthe Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi,UAE, 1-4 November. http://dx.doi.org/10.2118/136861-MS.
Harik, G.R., Lobo, F.G., and Goldberg, D.E. 1999. The compact geneticalgorithm. IEEE Trans. Evol. Comput. 3 (4): 287-297. http://dx.doi.org/10.1109/4235.797971.
Jahns, O.H. 1966. A Rapid Method for Obtaining a Two-Dimensional ReservoirDescription From Well Pressure Response Data. SPE J. 6 (4):315-327. SPE-1473-PA. http://dx.doi.org/10.2118/1473-PA.
Mohamed, L., Christie, M., and Demyanov, V. 2010. Comparison of StochasticSampling Algorithms for Uncertainty Quantification. SPE J. 15(1): 31-38. SPE-119139-PA. http://dx.doi.org/10.2118/119139-PA.
Mühlenbein, H. and Paaß, G. 1996. From recombination of genes to theestimation of distributions. Binary parameters. In Parallel Problem Solvingfrom Nature - PPSN IV, ed. H.M. Voigt, W. Ebeling, I. Rechenberg, and H.P.Schwefel, No. 1411, 178-187. Berlin: Lecture Notes in Computer Science,Springer.
Mühlenbein, H., Mahnig, T., and Rodriguez, A.O. 1999. Schemata,Distributions and Graphical Models in Evolutionary Optimization. Journal ofHeuristics 5 (2): 213-247. http://dx.doi.org/10.1023/A:1009689913453.
Pelikan, M., Goldberg, D.E., and Cant´u-Paz, E. 1999. BOA: The BayesianOptimization Algorithm. In Proceedings of the Genetic and EvolutionaryComputation Conference (GECCO '99), ed. W. Banzhaf, J. Daida, A.E. Eiben,et al., Vol. 1, 525-532. San Francisco, California: Morgan KaufmannPublishers.
Pelikan, M., Goldberg, D.E., and Cant´u-Paz, E. 2000. Linkage Problem,Distribution Estimation and Bayesian Networks. Evolutionary Computation8 (3): 311-340.
Petrovska, I. 2009. Estimation of distribution algorithms for reservoirhistory-matching optimisation. PhD thesis, Imperial College, London,UK.
Petrovska, I. and Carter, J.N. 2006. Estimation of distribution algorithmsfor history-matching. Paper presented at the 10th European Conference on theMathematics of Oil Recovery (ECMOR X), Amsterdam, 4-7 September.
Romero, C.E., Carter, J.N., Gringarten, A.C., and Zimmerman, R.W. 2000. AModified Genetic Algorithm for Reservoir Characterisation. Paper SPE 64765presented at the International Oil and Gas Conference and Exhibition in China,Beijing, 7-10 November. http://dx.doi.org/10.2118/64765-MS.
Sambridge, M.S. 1999. Geophysical inversion with a neighbourhoodalgorithm—II. Appraising the ensemble. Geophys. J. Int. 138 (3):727-746. http://dx.doi.org/10.1046/j.1365-246x.1999.00900.x.
Schulze-Riegert, R.W., Axmann, J.K., Haase, O., Rian, D.T., and You, Y.-L.2001. Optimization Methods for History Matching of Complex Reservoirs. PaperSPE 66393 presented at the SPE Reservoir Simulation Symposium, Houston, 11-14February. http://dx.doi.org/10.2118/66393-MS.
Schulze-Riegert, R.W., Krosche, M., Pajonk, O., and Mustafa, H. 2009. DataAssimilation Coupled to Evolutionary Algorithms—A Case Example in HistoryMatching. Paper SPE 125512 presented at the SPE/EAGE Reservoir Characterizationand Simulation Conference, Abu Dhabi, UAE, 19-21 October. http://dx.doi.org/10.2118/125512-MS.
Subbey, S., Christie, M.A., and Sambridge, M. 2003. A Strategy for RapidQuantification of Uncertainty in Reservoir Performance Prediction. Paper SPE79678 presented at the SPE Reservoir Simulation Symposium, Houston, 3-5February. http://dx.doi.org/10.2118/79678-MS.
Sultan, A.J., Ouenes, A., and Weiss, W.W. 1994. Automatic History Matchingfor an Integrated Reservoir Description and Improving Oil Recovery. Paper SPE27712 presented at the Permian Basin Oil and Gas Recovery Conference, Midland,Texas, USA, 16-18 March. http://dx.doi.org/10.2118/27712-MS.
Thomas, L.K., Hellums, L.J., and Reheis, G.M. 1972. A Nonlinear AutomaticHistory-Matching Technique for Reservoir Simulation Models. SPE J.12 (6): 508-514. SPE-3475-PA. http://dx.doi.org/10.2118/3475-PA.
Yang, C., Nghiem, L., Card, C., and Bremeier, M. 2007. Reservoir ModelUncertainty Quantification Through Computer-Assisted History Matching. PaperSPE 109825 presented at the SPE Annual Technical Conference and Exhibition,Anaheim, California, USA, 11-14 November. http://dx.doi.org/10.2118/109825-MS.