基于马尔科夫随机场的岩性识别方法
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
通过地震反演数据识别岩性,是地震反演的一项基本任务.由于不同岩性的弹性参数范围常常存在一定程度的重叠,所以给岩性识别带来了很大的困难.本文以叠前反演的弹性参数为基础,通过马尔科夫随机场(Markov Random Field简写为MRF)建立先验模型,按照解释好的测井资料,对不同岩性的弹性参数进行统计,得到计算所需的参数,在贝叶斯(Bayesian)框架下建立岩性分类的目标函数,达到岩性识别的目的.通过马尔科夫随机场建立先验模型,能够建立相邻点间的相互作用关系,得到横向上延续的岩性剖面.本文使用一个楔形模型和Marmousi Ⅱ模型对该方法进行了测试,结果表明,该方法有效可行.同时,本文通过加入误差的方法,检验了反演存在误差对识别结果的影响.
Lithologic discrimination by using parameters from seismic inversion is a basic task of seismic inversion.Because different lithologies usually have,to some extent,the similar elastic parameters,it is difficult to identify lithology.To solve this problem,lithologic discrimination method based on Markov random-field is applied.This method firstly builds a priori model through Markov random-field on the basis of elastic parameters of pre-stack inversion,and then obtains Gaussian distribution parameters of iterative computation by means of counting elastic parameters of different lithologies based on interpreted log data and creates objective function of lithologic discrimination under a Bayesian framework,and finally achieves the aim of lithologic discrimination.The priori model can establish interrelationships among adjacent points and obtain continuous lithologic sections.A wedge model and a Marmousi Ⅱ model are used to test the method.Results show that the method is feasible.Meanwhile,the influence of inversion error on lithologic discrimination accuracy is tested by adding error in this paper.
引文
[1]Ostrander W J.Plane-wave reflection coefficients for gassands at nonnormal angles of incidence.Geophysics,1984,49(10):1637-1648.
    [2]何又雄,姚姚.基于参考道的岩性识别与岩性剖面非线性反演.石油勘探与开发,2005,32(3):61-63.He Y X,Yao Y.Lithological profile non-linear inversion byreference trace-based lithological identification.PetroleumExploration and Development(in Chinese),2005,32(3):61-63.
    [3]李国福.多参数储层预测及流体识别方法研究[硕士论文].成都:成都理工大学,2011.Li G F.Multi-parameter reservoir prediction and fluididentification method research(in Chinese)[Master′s thesis].Chengdu:Chengdu University of Technology,2011.
    [4]Mukerji T,Avseth T,Mavko G,et al.Statistical rockphysics:Combining rock physics,information theory,andgeostatistics to reduce uncertainty in seismic reservoircharacterization.The Leading Edge,2001,20(3):313-319.
    [5]Eidsvik J,Avseth P,Omre H,et al.Stochastic reservoircharacterization using prestack seismic data.Geophysics,2004,69(4):978-993.
    [6]Bachrach R.Joint estimation of porosity and saturation usingstochastic rock-physics modeling.Geophysics,2006,71(5):O53-O63.
    [7]Larsen A L,Ulvmoen M,Omre H,et al.Bayesianlithology/fluid prediction and simulation on the basis of aMarkov-chain prior model.Geophysics,2006,71(5):R69-R78.
    [8]Gunning J,Glinsky M E.Detection of reservoir quality usingBayesian seismic inversion.Geophysics,2007,72(3):R37-R49.
    [9]Buland A,Kolbjrnsen O,Hauge R,et al.Bayesianlithology and fluid prediction from seismic prestack data.Geophysics,2008,73(3):C13-C21.
    [10]Eidsvik J,Omre H,Mukerji T,et al.Seismic reservoirprediction using Bayesian integration of rock physics andmarkov random fields:A North Sea example.The LeadingEdge,2002,21(3):290-294.
    [11]Spikes K,Mukerji T,Dvorkin J,et al.Probabilistic seismicinversion based on rock-physics models.Geophysics,2007,72(5):R87-R97.
    [12]Ulvmoen M,More H.Improved resolution in Bayesianlithology/fluid inversion from prestack seismic data and wellobservations:Part 1-Methodology.Geophysics,2010,75(2):R21-R35.
    [13]Ulvmoen M,More H,Buland A.Improved resolution inBayesian lithology/fluid inversion from prestack seismic dataand well observations:Part 2-Real case study.Geophysics,2010,75(2):B73-B82.
    [14]Ulvmoen M,Hammer H.Bayesian lithology/fluid inversion-comparison of two algorithms.Computational Geosciences,2010,14(2):357-367.
    [15]邓继新,王尚旭.基于统计岩石物理的含气储层饱和度与孔隙度联合反演.石油天然气学报,2009,31(1):48-53.Deng J X,Wang S X.Joint inversion of saturation andporosity in gas reservoirs based on statistical rock physics.Journal of Oil and Gas Technology(in Chinese),2009,31(1):48-53.
    [16]胡华锋.基于叠前道集的储层参数反演方法研究[硕士论文].青岛:中国石油大学(华东),2009.Hu H F.The Study of Petrophysical-Properties InversionBase on Pre-stack Seismic Data(in Chinese)[Master′sthesis].Qingdao:China University of Petroleum,2009.
    [17]Elfeki A,Dekking M.A Markov chain model for subsurfacecharacterization:Theory and applications.MathematicalGeology,2001,33(5):569-589.
    [18]Carle S F,Fogg G E.Modeling spatial variability with oneand multidimensional continuous-Lag Markov chains.MathematicalGeology,1997,29(7):891-918.
    [19]Weissmann G S,Fogg G E.Multi-scale alluvial fanheterogeneity modeled with transition probability geostatisticsin a sequence stratigraphic framework.Journal of Hydrology,1999,226(1-2):48-56.
    [20]Norberg T,Rosén L,Baran A,et al.On modeling discretegeological structures as Markov random fields.MathematicalGeology,2002,34(1):63-77.
    [21]李旭超,朱善安.图像分割中的马尔可夫随机场方法综述.中国图像图形学报,2007,12(5):789-798.Li X C,Zhu S A.A survey of the Markov random fieldmethod for image segmentation.Journal of Image andGraphics(in Chinese),2007,12(5):789-798.
    [22]Yuan S Y,Wang S X,Li G F.Random noise reduction usingBayesian inversion.Journal of Geophysics and Engineering,2012,9(1):60-68.
    [23]Pérez P.Markov random fields and images.CWI Quarterly,1998,11(4):413-437.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心