时移地震多属性智能分析在CO<sub>2</sub>驱油波及范围识别与圈定中的应用—以胜利油田G89区块为例

doi:10.3969/j.issn.2096-1693.2026.01.020

石油科学通报 ›› 2026, Vol. 11 ›› Issue (3): 836-849. doi: 10.3969/j.issn.2096-1693.2026.01.020

时移地震多属性智能分析在CO₂驱油波及范围识别与圈定中的应用—以胜利油田G89区块为例

刘浩辰¹^,²(), 刘钰铭¹^,²^,^*(), 曲志鹏¹^,²^,³, 张伟忠³, 张冰冰¹^,², 陈冠宇¹^,²

¹ 中国石油大学(北京) 油气资源与工程全国重点实验室，北京 102249
² 中国石油大学(北京)地球科学学院，北京 102249
³ 中国石化胜利油田物探研究院，东营 257001

收稿日期:2026-02-10 修回日期:2026-04-04 出版日期:2026-06-15 发布日期:2026-06-30
通讯作者: *刘钰铭(1983年—)，博士，教授，主要从事油气田开发地质学、油气藏表征与建模、油气地质大数据与人工智能技术等方面的教学与研究工作，liuym@cup.edu.cn。
作者简介:刘浩辰(2000年—)，博士研究生，主要从事CO₂驱油与封存地球物理监测与储层动态表征研究，2024315014@student.cup.edu.cn。
基金资助:
国家自然科学基金项目“盆缘过渡带坡度-流量双重控制下的辫状河成因机制与砂体构型模式”(42472205)

Application of intelligent time-lapse seismic multi-attribute analysis to the identification and delineation of CO₂ flooding sweep extent: Taking G89 area, Shengli oilfield as an example

LIU Haochen¹^,²(), LIU Yuming¹^,²^,^*(), QU Zhipeng¹^,²^,³, ZAHNG Weizhong³, ZAHNG Bingbing¹^,², CHEN Guanyu¹^,²

¹ State Key Laboratory of Petroleum Resources and Engineering, China University of Petroleum, Beijing 102249, China
² College of Geosciences, China University of Petroleum, Beijing 102249, China
³ Geophysical Research Institute of Shengli Oilfield Sinopec, Dongying 257001, China

Received:2026-02-10 Revised:2026-04-04 Online:2026-06-15 Published:2026-06-30
Contact: *liuym@cup.edu.cn

摘要/Abstract

摘要：

在“双碳”目标背景下，CO₂驱油作为兼具提高采收率与减排增效的优势开发方式，是推动油田高效开发与低碳转型的重要技术途径。准确识别CO₂运移通道与有效波及范围，是评价驱替效果和优化注采方案的关键。然而，常规生产动态分析及单一监测手段对地下CO₂运移过程与空间展布特征的刻画能力有限，难以满足精细表征需求。针对现有基于时移差异地震属性的识别方法易受噪声、非重复性误差及储层非均质性影响，导致异常响应离散、边界模糊等问题，本文提出一种基于时移地震多属性智能融合的CO₂驱油波及范围识别方法。基于时移地震资料构建差异体，优选振幅、相位及衰减等敏感差异属性，并结合模糊神经网络(FNN)，基于模糊规则实现多属性非线性融合，构建表征CO₂波及强弱的连续响应指标。结果表明，该方法能够有效抑制零散伪异常，提高预测结果的边界清晰度与空间连通性。时序对比显示，预测波及范围随注入推进由注气井周缘向外扩展，并沿上倾方向向构造高部位迁移聚集，圈定面积由2010年的约1.7 km²扩大至2022年的约2.6 km²。结合生产动态验证，高注气井与高产气井周缘普遍对应较强预测响应，表明该方法能够较准确反映CO₂驱油过程中的储层响应差异，为CO₂运移通道识别、波及范围定量表征及驱替效果评价提供了一种有效的地球物理技术途径。

关键词: 时移地震, 差异体, 时移差异属性分析, 多属性融合, 模糊神经网络, CO₂驱油

Abstract:

Under the “dual-carbon” goals, CO₂ flooding, as a development method that combines enhanced oil recovery with emission reduction and efficiency improvement, is an important technological approach to promote efficient oilfield development and low-carbon transformation. Accurate identification of CO₂ migration pathways and effective sweep extent are crucial for evaluating displacement effects and adjusting injection-production schemes. However, conventional production performance analysis and single monitoring methods have limited capability in characterizing the subsurface migration process and spatial distribution of CO₂, making it difficult to meet the demand for fine-scale characterization. To address the problems associated with existing identification methods based on time-lapse seismic difference attributes, which are susceptible to noise, non-repeatability errors, and reservoir heterogeneity and thus often lead to scattered anomalous responses and blurred boundaries, this study proposes a method for identifying the sweep extent of CO₂ flooding based on intelligent integration of time-lapse seismic multi-attributes. Difference volumes were constructed from time-lapse seismic data, and sensitive difference attributes, including amplitude, phase, and attenuation, were selected. Then, a fuzzy neural network (FNN) was introduced to perform nonlinear fusion of multiple attributes based on fuzzy rules, thereby constructing a continuous response indicator characterizing the intensity of CO₂ sweep. The results show that the proposed method can effectively suppress scattered false anomalies and improve the boundary clarity and spatial connectivity of the predicted results. Time-series comparison indicates that the predicted sweep extent expanded outward from the vicinity of injection wells as injection proceeded, and migrated upward along the up-dip direction toward structurally higher positions. The delineated sweep area increased from approximately 1.7 km² in 2010 to approximately 2.6 km² in 2022. Validation against production performance data further shows that strong predicted responses generally correspond to the vicinity of high gas-injection wells and high gas-production wells, indicating that the proposed method can more accurately reflect reservoir response differences during the CO₂ flooding process and provide an effective geophysical approach for identifying CO₂ migration pathways, quantitatively characterizing sweep extent, and evaluating displacement performance.

Key words: time-lapse seismic, difference volume, time-lapse difference attribute analysis, multi-attribute fusion, fuzzy neural network, CO₂ flooding

中图分类号:

刘浩辰, 刘钰铭, 曲志鹏, 张伟忠, 张冰冰, 陈冠宇. 时移地震多属性智能分析在CO₂驱油波及范围识别与圈定中的应用—以胜利油田G89区块为例[J]. 石油科学通报, 2026, 11(3): 836-849.

LIU Haochen, LIU Yuming, QU Zhipeng, ZAHNG Weizhong, ZAHNG Bingbing, CHEN Guanyu. Application of intelligent time-lapse seismic multi-attribute analysis to the identification and delineation of CO₂ flooding sweep extent: Taking G89 area, Shengli oilfield as an example[J]. Petroleum Science Bulletin, 2026, 11(3): 836-849.

https://sykxtb.cup.edu.cn/CN/Y2026/V11/I3/836

图/表 10

图1 研究区地理位置及储层地质分层信息图

Fig. 1 Geographical location of the study area and stratigraphic subdivision of the reservoir

图2 G89区块基线与监测三维地震资料NRMS平面分布及频数分布

Fig. 2 NRMS map and frequency distribution of baseline and monitor 3D seismic data in the G89 block

图3 G89区块三期三维地震资料代表性属性平面分布对比

Fig. 3 Representative attribute maps of baseline and monitor 3D seismic datasets in the G89 block

图4 G89区块两阶段时移差异地震体代表性差异属性平面分布对比

Fig. 4 Representative difference-attribute maps of two-stage time-lapse seismic difference volumes in the G89 block

表1 2010—2022年差异属性筛选统计量

Table 1 Statistics of selected difference attributes from 2010 to 2022

属性类型	盖层均值μ_s	盖层标准差σ_s	目标层均值μ_t	目标层标准差σ_t	异常幅值比Ra	层段区分度D
∆最大绝对振幅	0.564	0.096	0.567	0.112	0.031	0.021
∆平均瞬时频率	0.210	0.080	0.226	0.100	0.001	0.128
∆平均瞬时相位	0.379	0.212	0.554	0.161	0.823	0.656
∆均方根振幅	0.169	0.122	0.251	0.153	0.674	0.421
∆能量半衰时	0.043	0.068	0.086	0.101	0.630	0.352
∆平均反射强度	0.210	0.127	0.240	0.143	0.239	0.159
∆总振幅	0.550	0.099	0.594	0.111	0.442	0.294
∆吸收衰减	0.511	0.110	0.427	0.112	0.760	0.533

表2 1992—2010年差异属性筛选统计量

Table 2 Statistics of selected difference attributes from 1992—2010

属性类型	盖层均值μ_s	盖层标准差σ_s	目标层均值μ_t	目标层标准差σ_t	异常幅值比Ra	层段区分度D
∆最大绝对振幅	0.408	0.093	0.447	0.124	0.419	0.250
∆平均瞬时频率	0.401	0.101	0.417	0.111	0.160	0.108
∆平均瞬时相位	0.584	0.340	0.353	0.200	0.680	0.586
∆均方根振幅	0.161	0.120	0.253	0.150	0.767	0.478
∆能量半衰时	0.039	0.063	0.049	0.071	0.149	0.098
∆平均反射强度	0.179	0.110	0.207	0.122	0.256	0.172
∆总振幅	0.399	0.096	0.421	0.118	0.228	0.144
∆吸收衰减	0.488	0.110	0.598	0.120	1.053	0.674

图5 G89区块沙四段纯下次亚段目标层位两阶段优选差异地震属性平面分布对比

Fig. 5 Comparison of selected time-lapse seismic difference-attribute maps at the target horizon for two stages in the Gao89 block

图6 模糊神经网络结构

Fig. 6 Schematic diagram of the fuzzy neural network architecture

图7 G89区块沙四段纯下次亚段CO₂驱油波及强度的FNN多属性预测结果

Fig. 7 FNN-based multi-attribute prediction of CO₂ flooding sweep intensity in the pure lower sub-member of the Es₄ Member, G89 block

图8 G89井区CO₂驱油波及范围

Fig. 8 CO₂ flooding sweep extent in the G89 area

参考文献 40

[1]	DAVOODI S, AL-SHARGABI M, WOOD D A, et al. Carbon dioxide sequestration through enhanced oil recovery: A review of storage mechanisms and technological applications[J]. Fuel, 2024, 366: 131313. doi: 10.1016/j.fuel.2024.131313 URL
[2]	张成龙, 胡丽莎, 牛兆轩, 等. CO₂-EOR在低渗透滩坝砂油藏的应用初探——以高89为例[J]. 科学技术与工程, 2023, 23(15): 6393-6401.
	[ZHANG C L, HU L S, NIU Z X, et al. Application of CO₂-EOR in low-permeability beach-bar sand reservoir: Taking Shengli oilfield Gao89 block as an example[J]. Science Technology and Engineering, 2023, 23(15): 6393-6401.]
[3]	秦积舜, 李永亮, 吴德斌, 等. CCUS全球进展与中国对策建议[J]. 油气地质与采收率, 2020, 27(1): 20-28.
	[QIN J S, LI Y L, WU D B, et al. CCUS global progress and China’s policy suggestions[J]. Petroleum Geology and Recovery Efficiency, 2020, 27(1): 20-28.]
[4]	庞敏, 张益畅. 我国发展二氧化碳驱油技术的路径思考[J]. 西南石油大学学报(社会科学版), 2024, 26(4): 1-9. doi: 10.11885/j.issn.1674 5094.2023.11.17.01
	[PANG M, ZHANG Y C. The path of developing CO₂ flooding technology in China[J]. Journal of Southwest Petroleum University (Social Sciences Edition), 2024, 26(4): 1-9.]
[5]	李阳, 黄文欢, 金勇, 等. 双碳愿景下中国石化不同油藏类型CO₂驱提高采收率技术发展与应用[J]. 油气藏评价与开发, 2021, 11(6): 793-804, 790.
	[LI Y, HUANG W H, JIN Y, et al. Different reservoir types of CO₂ flooding in Sinopec EOR technology development and application under “dual carbon” vision[J]. Petroleum Reservoir Evaluation and Development, 2021, 11(6): 793-804, 790.]
[6]	杨谦, 梁金强, 庹林峰. CO₂驱油技术的发展与挑战[J]. 精细石油化工进展, 2025, 26(5): 22-27.
	[YANG Q, LIANG J Q, TUO L F. Development and challenge of CO₂ flooding technology[J]. Advances in Fine Petrochemicals, 2025, 26(5): 22-27.]
[7]	张志超. 低渗透油藏CO₂驱油及封存一体化研究及泄漏评价[D]. 大庆: 东北石油大学, 2025.
	[ZHANG Z C. Study on the integration of CO₂ flooding oil and storage and evaluation of the CO₂ leakage risk in low-permeable reservoirs[D]. Daqing: Northeast Petroleum University, 2025.]
[8]	赵辉, 陈子风, 盛广龙, 等. CO₂驱油封存优势窜流识别及调控优化[J]. 科学通报, 2025, 70(13): 1977-1984.
	[ZHAO H, CHEN Z F, SHENG G L, et al. Predominant channeling identification and regulation optimization of CO₂ flooding storage[J]. Chinese Science Bulletin, 2025, 70(13): 1977-1984.]
[9]	WANG H F, WANG Z, MA J F, et al. Effective pressure prediction from 4D seismic AVO data during CO₂-EOR and storage[J]. International Journal of Greenhouse Gas Control, 2022, 113: 103525. doi: 10.1016/j.ijggc.2021.103525 URL
[10]	李春雷, 赵程, 谢涛, 等. 渤海湾盆地疏松砂岩岩石物理建模及时移地震可行性分析[J]. 石油科学通报, 2025, 10(4): 681-694
	[LI C L, ZHAO C, XIE T, et al. Rock physics modeling and time-lapse seismic feasibility analysis of unconsolidated sandstones in the Bohai Bay Basin[J]. Petroleum Science Bulletin, 2025, 10(4): 681-694.]
[11]	李文瑾, 李景叶, 王永平. 基于L1-2范数的时移波阻抗反演方法[J]. 石油科学通报, 2024, 9(6): 921-930.
	[LI W J, LI J Y, WANG Y P. Time-lapse impedance inversion based on L1-2 norm[J]. Petroleum Science Bulletin, 2024, 9(6): 921-930.]
[12]	MEADOWS M A, COLE S P. 4D seismic modeling and CO₂ pressure-saturation inversion at the Weyburn field, Saskatchewan[J]. International Journal of Greenhouse Gas Control, 2013, 16: S103-S117. doi: 10.1016/j.ijggc.2013.01.030 URL
[13]	FAWAD M, MONDOL N H. Monitoring geological storage of CO₂ using a new rock physics model[J]. Scientific Reports, 2022, 12: 297. doi: 10.1038/s41598-021-04400-7
[14]	WIPKI M, IVANOVA A, LIEBSCHER A, et al. Monitoring concept for CO₂ storage at the Ketzin pilot site, Germany: Post-injection continuation towards transfer of liability[J]. Energy Procedia, 2016, 97: 348-355. doi: 10.1016/j.egypro.2016.10.017 URL
[15]	MARTENS S, LIEBSCHER A, MÖLLER F, et al. Progress report on the first European on-shore CO₂ storage site at Ketzin (Germany): Second year of injection[J]. Energy Procedia, 2011, 4: 3246-3253. doi: 10.1016/j.egypro.2011.02.243 URL
[16]	HEMA G, MAURYA S, KANT R, et al. Enhancement of CO₂ monitoring in the Sleipner field (North Sea) using seismic inversion based on simulated annealing of time-lapse seismic data[J]. Marine and Petroleum Geology, 2024, 167: 106962. doi: 10.1016/j.marpetgeo.2024.106962 URL
[17]	ARTS R, EIKEN O, CHADWICK A, et al. Monitoring of CO₂ injected at Sleipner using time-lapse seismic data[J]. Energy, 2004, 29(9): 1383-1392. doi: 10.1016/j.energy.2004.03.072 URL
[18]	SINGH A P, MAURYA S P, KANT R, et al. Implementing 4D seismic inversion based on linear programming techniques for CO₂ monitoring at the Sleipner field CCS site in the North Sea, Norway[J]. Acta Geophysica, 2025, 73(1): 271-293. doi: 10.1007/s11600-024-01376-6
[19]	CHADWICK R A, NOY D, ARTS R, et al. Latest time-lapse seismic data from Sleipner yield new insights into CO₂ plume development[J]. Energy Procedia, 2009, 1(1): 2103-2110. doi: 10.1016/j.egypro.2009.01.274 URL
[20]	CHADWICK R A, MARCHANT B P, WILLIAMS G A. CO₂ storage monitoring: Leakage detection and measurement in subsurface volumes from 3D seismic data at Sleipner[J]. Energy Procedia, 2014, 63: 4224-4239. doi: 10.1016/j.egypro.2014.11.458 URL
[21]	EIKEN O, RINGROSE P, HERMANRUD C, et al. Lessons learned from 14 years of CCS operations: Sleipner, in Salah and Snøhvit[J]. Energy Procedia, 2011, 4: 5541-5548. doi: 10.1016/j.egypro.2011.02.541 URL
[22]	程丽媛, 贾小宝, 刘燕海. 基于储层多参数同时预测的煤层CO₂运移监测方法研究[J]. 石油物探, 2025, 64(5): 970-978.
	[CHENG L Y, JIA X B, LIU Y H. Monitoring method of CO₂ migration in coal seam based on simultaneous prediction of reservoir multi-parameters[J]. Geophysical Prospecting for Petroleum, 2025, 64(5): 970-978.]
[23]	张伟忠, 谭明友, 张云银, 等. S油田高89区块基于AVO属性的CO₂驱油波及范围地震监测[J]. 石油物探, 2019, 58(2): 285-291. doi: 10.3969/j.issn.1000-1441.2019.02.014
	[ZHANG W Z, TAN M Y, ZHANG Y Y, et al. Seismic detection of the CO₂ flooding extent based on the AVO attributes in the G89 area in S oilfield, China[J]. Geophysical Prospecting for Petroleum, 2019, 58(2): 285-291.]
[24]	张军华, 杨梅, 陈永芮, 等. CO₂驱油地震监测技术的研究现状与进展[J]. 石油地球物理勘探, 2025, 60(2): 532-543.
	[ZHANG J H, YANG M, CHEN Y R, et al. Research status and progress of seismic monitoring technology for CO₂ flooding[J]. Oil Geophysical Prospecting, 2025, 60(2): 532-543.]
[25]	芮拥军, 尚新民. 胜利油田非一致性时移地震关键技术探索与实践[J]. 物探与化探, 2021, 45(6): 1439-1447.
	[RUI Y J, SHANG X M. Exploration and practice of non-uniform time-lapse seismic key technology in Shengli Oilfield[J]. Geophysical and Geochemical Exploration, 2021, 45(6): 1439-1447.]
[26]	刘巍, 雷霄, 隋波, 等. 压缩感知在非重复时移地震处理中应用研究[J]. 海洋科学, 2025, 49 (8): 97-107.
	[LIU W, LEI X, SUI B, et al. Application of compressed sensing in multicomponent and towed streamer data in time-shift seismic processing[J]. Marine Sciences, 2025, 49 (8): 97-107.]
[27]	侯斯允, 李景叶, 耿伟恒, 等. 基于微分超拉普拉斯块约束的改进弹性阻抗反演方法[J]. 石油科学通报, 2024, 9(4): 549-562.
	[HOU S Y, LI J Y, GENG W H, et al. A modified inversion method for elastic impedance based on the differentiable hyper-Laplacian blocky constraint[J]. Petroleum Science Bulletin, 2024, 9(4): 549-562.]
[28]	汤国松, 李皓, 梁兵, 等. 空间结构正则化多道稀疏脉冲反褶积[J]. 石油科学通报, 2024, 9(6): 911-920.
	[TANG G S, LI H, LIANG B, et al. Spatially structured regularization multichannel sparse pulse deconvolution[J]. Petroleum Science Bulletin, 2024, 9(6): 911-920.]
[29]	BARRELA E, BERTHET P, TRANI M, et al. Four-dimensional history matching using ES-MDA and flow-based distance-to-front measurement[J]. Energies, 2023, 16(24): 7984. doi: 10.3390/en16247984 URL
[30]	JERVIS M, BAKULIN A, SMITH R. Making time-lapse seismic work in a complex desert environment for CO₂ EOR monitoring: Design and acquisition[J]. The Leading Edge, 2018, 37(8): 598-606. doi: 10.1190/tle37080598.1 URL
[31]	王鑫, 刘惠民, 张鹏飞, 等. 渤海湾盆地济阳坳陷博兴洼陷沙四上亚段页岩层系流体包裹体特征及其油气地质意义[J]. 天然气地球科学, 2026, 37(3): 518-527. doi: 10.11764/j.issn.1672-1926.2025.11.005
	[WANG X, LIU H M, ZHANG P F, et al. The development characteristics of fluid inclusions and its petroleum geological significancein the upper fourth member of Shahejie Formation in Boxing subsag of Jiyang Depression in Bohai Bay Basin[J]. Natural Gas Geoscience, 2026, 37(3): 518-527.] doi: 10.11764/j.issn.1672-1926.2025.11.005
[32]	谭明友, 曲志鹏, 张云银, 等. 二氧化碳驱油地震监测评价方法研究[M]. 北京: 科学出版社, 2017.
	[TAN M Y, QU Z P, ZHANG Y Y, et al. 4D seismic monitoring of Carbon Dioxide flooding[M]. Beijing: Science Press, 2017.]
[33]	MA J F, YANG Y, WANG H F, et al. How much CO₂is stored and verified through CCS/CCUS in China[J]. Energy Procedia, 2018, 154: 60-65. doi: 10.1016/j.egypro.2018.11.011 URL
[34]	WANG H F, MA J F, LI L, et al. Time-lapse seismic analysis for Gao89 area of CO₂-EOR project in SINOPEC Shengli Oilfield, China[J]. Energy Procedia, 2017, 114: 3980-3988. doi: 10.1016/j.egypro.2017.03.1530 URL
[35]	王喆, 李琳, 马劲风, 等. CO₂驱油与封存过程中时移地震纹理属性分析——以胜利油田G89区块为例[J]. 地球物理学进展, 2020, 35(4): 1481-1488.
	[WANG Z, LI L, MA J F, et al. Time-lapse seismic texture analysis in CO₂ -EOR and storage: G89 area, Sinopec Shengli oilfield, China[J]. Progress in Geophysics, 2020, 35(4): 1481-1488.]
[36]	陈筱, 杨希濮, 肖鹏, 等. 时移地震技术在油藏监测中的应用——以西非深水扇A油田为例[J]. 石油物探, 2023, 62(3): 538-547. doi: 10.12431/issn.1000-1441.2023.62.03.015
	[CHEN X, YANG X P, XIAO P, et al. Application of time-lapse seismic technology in deep water turbidite reservoir monitoring: A case study of the Deepwater Fan A oilfield in West Africa[J]. Geophysical Prospecting for Petroleum, 2023, 62(3): 538-547.]
[37]	岳大力, 李伟, 杜玉山, 等. 河流相储层地震属性优选与融合方法综述[J]. 地球科学, 2022, 47(11): 3929-3943.
	[YUE D L, LI W, DU Y S, et al. Review on optimization and fusion of seismic attributes for fluvial reservoir characterization[J]. Earth Science, 2022, 47(11): 3929-3943.]
[38]	LAKENS D. Calculating and reporting effect sizes to facilitate cumulative science: A practical primer for t-tests and ANOVAs[J]. Frontiers in Psychology, 2013, 4: 863. doi: 10.3389/fpsyg.2013.00863 pmid: 24324449
[39]	ZHAO J W, CHANG D J, CAO B, et al. Multiobjective evolution of the deep fuzzy rough neural network[J]. IEEE Transactions on Fuzzy Systems, 2025, 33(1): 242-254. doi: 10.1109/TFUZZ.2024.3397728 URL
[40]	LIU Y, LIU Y Q, SHAO Q, et al. A novel neuro-fuzzy learning algorithm for first-order Takagi-Sugeno fuzzy model: Caputo fractional-order gradient descent method[J]. International Journal of Fuzzy Systems, 2024, 26(8): 2616-2631. doi: 10.1007/s40815-024-01750-y

选择文件类型/文献管理软件名称

选择包含的内容

时移地震多属性智能分析在CO₂驱油波及范围识别与圈定中的应用—以胜利油田G89区块为例

Application of intelligent time-lapse seismic multi-attribute analysis to the identification and delineation of CO₂ flooding sweep extent: Taking G89 area, Shengli oilfield as an example

RichHTML

PDF

可视化