深水扇储层物性多属性反演方法研究——基于“步聪法”进行敏感地震属性组合优选
|
摘要
深水扇储层砂泥岩纵波阻抗的动态值域范围常常相互重叠,采用常规纵波阻抗反演方法无法进行精确的储层物性描述。为解决此类问题,在探讨"步聪法"优选敏感地震属性组合的基础上,利用概率神经网络算法建立储层物性参数与敏感地震属性组合之间的非线性关系,实现了储层物性参数的直接反演。该方法在A油田应用取得了好的效果,成功解决了深水扇储层砂泥岩纵波阻抗动态值域范围相互重叠情况下的岩性识别及储层物性描述问题。
The dynamic ranges of P-wave impedance for shale and sandstone will frequently overlap in deep-water fan systems,and their reservoir petrophysics can not be accurately described with the conventional method of P-wave impedance inversion.In order to solve this problem,through a discussion of optimizing a combination of sensitive seismic attributes by "the step-wise method",some nonlinear relationships between the reservoir petrophysics and the optimal combination of sensitive seismic attributes were established with an algorithm of probabilistic neural network,resulting in a direct inversion of reservoir petrophysics parameters.An application of this method in oilfield A has led to good effects,in which the problems of lithologic identification and reservoir petrophysics description are successfully solved in the deep-water fan systems where the dynamic ranges of P-wave impedance for shale and sandstone may overlap.
The dynamic ranges of P-wave impedance for shale and sandstone will frequently overlap in deep-water fan systems,and their reservoir petrophysics can not be accurately described with the conventional method of P-wave impedance inversion.In order to solve this problem,through a discussion of optimizing a combination of sensitive seismic attributes by "the step-wise method",some nonlinear relationships between the reservoir petrophysics and the optimal combination of sensitive seismic attributes were established with an algorithm of probabilistic neural network,resulting in a direct inversion of reservoir petrophysics parameters.An application of this method in oilfield A has led to good effects,in which the problems of lithologic identification and reservoir petrophysics description are successfully solved in the deep-water fan systems where the dynamic ranges of P-wave impedance for shale and sandstone may overlap.
引文
[1]CHEN Quincy,SIDNEY S.Seismic attribute technology forreservoir forecasting and monnitoring[J].The Leading Edge,1997,16(5):445-456.
[2]姜秀清.储层地震属性优化及属性体综合解释[D].广州:中国科学院研究生院,2006.
[3]彭路平,王一刚,刘划一.地震多属性对碳酸盐岩有利储集层及含气性预测[J].中国石油勘探,1999,4(3):53-56.
[4]COOKE D,BALL V.What is the best seismic attribute forquantitative seismic reservoir characterization[C]SEG6 9th Expanded Abstracts,1999.
[5]陈遵德.储层地震属性优化方法[M].北京:石油工业出版社,1996.
[6]罗德江.地震属性优化和神经网络在储层预测中的应用[D].成都:成都理工大学,2006.
[7]张志国.人工神经网络及其在地学中的应用研究[D].吉林:吉林大学,2006.
[8]蔡曲林.基于概率神经网络的模式识别[D].长沙:国防科学技术大学,2005.
[9]刘文岭.牛彦良,李刚,等.多信息储层预测地震属性提取与有效性分析方法[J].石油物探,2002,41(1):100-106.
[2]姜秀清.储层地震属性优化及属性体综合解释[D].广州:中国科学院研究生院,2006.
[3]彭路平,王一刚,刘划一.地震多属性对碳酸盐岩有利储集层及含气性预测[J].中国石油勘探,1999,4(3):53-56.
[4]COOKE D,BALL V.What is the best seismic attribute forquantitative seismic reservoir characterization[C]SEG6 9th Expanded Abstracts,1999.
[5]陈遵德.储层地震属性优化方法[M].北京:石油工业出版社,1996.
[6]罗德江.地震属性优化和神经网络在储层预测中的应用[D].成都:成都理工大学,2006.
[7]张志国.人工神经网络及其在地学中的应用研究[D].吉林:吉林大学,2006.
[8]蔡曲林.基于概率神经网络的模式识别[D].长沙:国防科学技术大学,2005.
[9]刘文岭.牛彦良,李刚,等.多信息储层预测地震属性提取与有效性分析方法[J].石油物探,2002,41(1):100-106.
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