基于增强型生成对抗网络地质建模框架的多模式非平稳储层随机模拟

宋随宏; MUKERJI Tapan; SCHEIDT Celine; ALQASSAB Hisham M.; FENG Man

doi:10.11698/PED.202500125SD

石油勘探与开发 >

0 20260213 - 20260213

DOI: https://doi.org/10.11698/PED.202500125SD

基于增强型生成对抗网络地质建模框架的多模式非平稳储层随机模拟

宋随宏 ,
MUKERJI Tapan ,
SCHEIDT Celine ,
ALQASSAB Hisham M. ,
FENG Man

展开

1.斯坦福大学能源科学与工程系,加州 94305,美国;
2.埃克森美孚技术与工程公司,德克萨斯州 77389,美国

宋随宏（1990-）,男,陕西延安人,博士,斯坦福大学研究科学家（Research Scientist）,主要从事人工智能、储层地质建模、渗流模拟和AI大模型研发与应用方面的工作。地址：Stanford University, 367 Panama St, Stanford, CA 94305, USA。E-mail：suihong@stanford.edu

收稿日期: 2025-11-30

修回日期: 2026-01-19

网络出版日期: 2026-01-22

收起

Geomodelling of multi-scenario non-stationary reservoirs with enhanced GANSim

SONG Suihong ,
MUKERJI Tapan ,
SCHEIDT Celine ,
ALQASSAB Hisham M. ,
FENG Man

Expand

1. Department of Energy Science and Engineering, Stanford University, 367 Panama St, Stanford, CA 94305, USA;
2. ExxonMobil Technology & Engineering Company, Spring, Texas, 77389 USA

Received date: 2025-11-30

Revised date: 2026-01-19

Online published: 2026-01-22

Fold

摘要

为了将GANSim（基于生成对抗网络的直接条件化地质建模框架）用于多地质模式、非平稳性储层地质建模,并克服该方法易于忽略占据单网格的井点条件数据进而导致井周沉积相模拟结果局部断连的局限,提出“增强型GANSim”建模框架,以多模式非平稳浊积扇储层地质建模为例,验证增强型GANSim的有效性。针对可能存在多种地质模式的储层,提出两种GANSim地质建模流程：①训练一个覆盖所有可能地质模式的综合GANSim模型;②先进行地质模式证伪,然后针对未被证伪的地质模式训练GANSim模型。在此基础上,进行局部判别器的结构设计,以提升井周沉积相的连续性。浊积扇储层建模结果表明,两种GANSim建模流程均能产生与期望地质规律和条件数据都吻合的非平稳性沉积相地质模型实现,同时井周沉积相不连续的问题也得以解决。与多点地质统计学方法（SNESIM）相比,GANSim表现出卓越的储层规律复现能力和建模效率,尽管GANSim的训练耗时较长,但是一旦训练完成,即可用于任何具有相似地质结构、任意规模储层的地质建模,建模速度约为SNESIM的1 000倍。

关键词： 储层地质建模; 生成对抗网络; 增强型GANSim; 模式证伪; 非平稳性储层; 浊积扇

本文引用格式

宋随宏 , MUKERJI Tapan , SCHEIDT Celine , ALQASSAB Hisham M. , FENG Man . 基于增强型生成对抗网络地质建模框架的多模式非平稳储层随机模拟[J]. 石油勘探与开发, 0 : 20260213 -20260213 . DOI: 10.11698/PED.202500125SD

Abstract

GANSim is a generative adversarial networks (GANs)-based direct conditional geomodelling framework. To extend GANSim to multi-scenario, non-stationary reservoir geomodelling, and to address its tendency to overlook single-grid well conditioning data that can cause local facies disconnections around wells, an enhanced GANSim framework is proposed. The effectiveness of the enhanced GANSim is validated using a 3D multi-scenario, non-stationary turbidite reservoir as an example. For reservoirs that may involve multiple geological scenarios, two GANSim geomodelling workflows are proposed: (1) training a comprehensive GANSim model that covers all possible geological scenarios; and (2) first performing geological scenario falsification and then training GANSim models only for the unfalsified scenarios. On this basis, a local discriminator architecture is designed to improve facies continuity around wells. The modelling results show that both workflows can generate non-stationary facies models that conform to expected geological patterns and honor conditioning data, and the facies discontinuity issue around wells is effectively resolved. Compared with multipoint geostatistical methods (SNESIM), GANSim exhibits superior capability in reproducing reservoir geological patterns and modelling efficiency. Although GANSim requires a longer training time, once training is completed, it can be applied to geomodelling reservoirs of arbitrary scale with similar geological structures, achieving modelling speeds approximately 1000 times faster than SNESIM.

Key words： reservoir geomodelling; generative adversarial networks (GANs); enhanced GANSim; scenario falsification; non-stationary reservoirs; turbidite

参考文献

[1] DEUTSCH C V.Geostatistical reservoir modeling[M]. New York: Oxford University Press, 2002.
[2] MARIETHOZ G, CAERS J.Multiple‐point geostatistics: Stochastic modeling with training images[M]. Chichester: John Wiley & Sons Inc., 2014.
[3] SONG S H, HUANG J Y, MUKERJI T. Generative geomodelling: Deep learning vs. geostatistics[EB/OL]. (2025-06-20)[2025-10-20]. https://doi.org/10.31223/X5B732.
[4] GOODFELLOW I J, POUGET-ABADIE J, MIRZA M, et al.Generative adversarial nets[C]//GHAHRAMANI Z, WELLING M, CORTES C, et al. Advances in Neural Information Processing Systems 27. Red Hook, NY: Curran Associates, Inc., 2016: 2672-2680.
[5] HO J, JAIN A N, ABBEEL P.Denoising diffusion probabilistic models[C]//LAROCHELLE H, RANZATO M, HADSELL R, et al. Advances in Neural Information Processing Systems 33. Red Hook, NY: Curran Associates, Inc., 2020: 6840-6851.
[6] ZHANG T F, TILKE P, DUPONT E, et al.Generating geologically realistic 3D reservoir facies models using deep learning of sedimentary architecture with generative adversarial networks[J]. Petroleum Science, 2019, 16(3): 541-549.
[7] YANG Z X, CHEN Q Y, CUI Z S, et al.Automatic reconstruction method of 3D geological models based on deep convolutional generative adversarial networks[J]. Computational Geosciences, 2022, 26(5): 1135-1150.
[8] SONG S H, MUKERJI T, HOU J G, et al. GANSim-3D for conditional geomodeling: Theory and field application[J]. Water Resources Research, 2022, 58(7): e2021WR031865.
[9] SONG S H, ZHANG D X, MUKERJI T, et al.GANSim-surrogate: An integrated framework for stochastic conditional geomodelling[J]. Journal of Hydrology, 2023, 620(Part B): 129493.
[10] FAN W Y, LIU G, CHEN Q Y, et al.Geological model automatic reconstruction based on conditioning Wasserstein generative adversarial network with gradient penalty[J]. Earth Science Informatics, 2023, 16(3): 2825-2843.
[11] HU F, WU C L, SHANG J W, et al.Multi-condition controlled sedimentary facies modeling based on generative adversarial network[J]. Computers & Geosciences, 2023, 171: 105290.
[12] CUI Z S, CHEN Q Y, LUO J, et al. Characterizing subsurface structures from hard and soft data with multiple-condition fusion neural network[J]. Water Resources Research, 2024, 60(11): e2024WR038170.
[13] ALQASSAB H M, FENG M, BECKER J A, et al.MAGCS: Machine assisted geologic carbon storage[R]. SPE 222120-MS, 2024.
[14] LALOY E, HÉRAULT R, JACQUES D, et al. Training-image based geostatistical inversion using a spatial generative adversarial neural network[J]. Water Resources Research, 2018, 54(1): 381-406.
[15] SONG S H, MUKERJI T, HOU J G.Geological facies modeling based on progressive growing of generative adversarial networks (GANs)[J]. Computational Geosciences, 2021, 25(3): 1251-1273.
[16] MO S X, ZABARAS N, SHI X Q, et al. Integration of adversarial autoencoders with residual dense convolutional networks for estimation of non-Gaussian hydraulic conductivities[J]. Water Resources Research, 2020, 56(2): e2019WR026082.
[17] MOSSER L, DUBRULE O, BLUNT M J.Stochastic seismic waveform inversion using generative adversarial networks as a geological prior[J]. Mathematical Geosciences, 2020, 52(1): 53-79.
[18] NESVOLD E, MUKERJI T. Simulation of fluvial patterns with GANs trained on a data set of satellite imagery[J]. Water Resources Research, 2021, 57(5): e2019WR025787.
[19] SONG S H, MUKERJI T, HOU J G.GANSim: Conditional facies simulation using an improved progressive growing of generative adversarial networks (GANs)[J]. Mathematical Geosciences, 2021, 53(7): 1413-1444.
[20] SONG S H, MUKERJI T, HOU J G.Bridging the gap between geophysics and geology with generative adversarial networks[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 5902411.
[21] HU X, SONG S H, HOU J G, et al. Stochastic modeling of thin mud drapes inside point bar reservoirs with ALLUVSIM-GANSim[J]. Water Resources Research, 2024, 60(6): e2023WR035989.
[22] SONG S H, MUKERJI T, ZHANG D X.Physics-informed multi-grid neural operator: Theory and an application to porous flow simulation[J]. Journal of Computational Physics, 2025, 520: 113438.
[23] BORG I, GROENEN P J F. Modern multidimensional scaling: Theory and applications[M]. 2nd ed. New York: Springer, 2005.
[24] SCHEIDT C, TAHMASEBI P, PONTIGGIA M, et al.Updating joint uncertainty in trend and depositional scenario for reservoir exploration and early appraisal[J]. Computational Geosciences, 2015, 19(4): 805-820.
[25] PARK H, SCHEIDT C, FENWICK D, et al.History matching and uncertainty quantification of facies models with multiple geological interpretations[J]. Computational Geosciences, 2013, 17(4): 609-621.
[26] SCHEIDT C, JEONG C, MUKERJI T, et al.Probabilistic falsification of prior geologic uncertainty with seismic amplitude data: Application to a turbidite reservoir case[J]. Geophysics, 2015, 80(5): M89-M12.
[27] JOURNEL A G.Combining knowledge from diverse sources: An alternative to traditional data independence hypotheses[J]. Mathematical Geology, 2002, 34(5): 573-596.
[28] GULRAJANI I, AHMED F, ARJOVSKY M, et al.Improved training of Wasserstein GANs[C]//VON LUXBURG U, GUYON I, BENGIO S, et al. Proceedings of the 31st International Conference on Neural Information Processing Systems. Red Hook, NY: Curran Associates Inc., 2017: 5769-5779.
[29] MCHARGUE T R, HODGSON D M, SHELEF E.Architectural diversity of submarine lobate deposits[J]. Frontiers in Earth Science, 2021, 9: 697170.
[30] DEPTUCK M E, SYLVESTER Z, PIRMEZ C, et al.Migration-aggradation history and 3-D seismic geomorphology of submarine channels in the Pleistocene Benin-major Canyon, western Niger Delta slope[J]. Marine and Petroleum Geology, 2007, 24(6/7/8/9): 406-433.
[31] MCHARGUE T, PYRCZ M J, SULLIVAN M D, et al.Architecture of turbidite channel systems on the continental slope: Patterns and predictions[J]. Marine and Petroleum Geology, 2011, 28(3): 728-743.
[32] AVSETH P, MUKERJI T, MAVKO G.Quantitative seismic interpretation: Applying rock physics tools to reduce interpretation risk[M]. Cambridge: Cambridge University Press, 2005.
[33] KINGMA D P, BA J. Adam: A method for stochastic optimization[DB/OL]. (2014-12-22)[2025-11-12]. https://doi.org/10. 48550/arXiv.1412.6980.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献