地震属性模式聚类预测储层物性参数
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
地震模式聚类预测技术是一种模式识别技术,在油气储层预测中有广泛的应用。其应用成功的关键在于:1所用地震资料的质量满足高信噪比、高分辨率及高保真度的要求;2建立准确的地质模型(综合框架) ,为神经网络的训练提供好的学习样本;3选用多层感知器神经网络。在上述第二个环节中,涉及储层层段划分上应力求按岩性特征细化,然后按测井细分层数据对地震信息进行准确标定,建立地震属性信息与地质、井信息的联系。文中列举了两个油田的应用实例,进一步证实了此方法的应用效果。
Seismic pattern cluster prediction technique is a pattern recognition technique,which has wide application in the prediction of oil/gas reservoir.The key of successful application lies as follows:①the quality of seismic data used meet the needs of high S/N ratio,high resolution and hi-fi;②building up correct geological model (integrative framework) so that can provide good study sample for neural network training;③selecting neural network with multi-layer sensor.In above-mentioned second link,in concern with formation-partitioning of reservoir it must make effort to refine according to lithology characters,then correctly labele the seismic information according to layer-refined data by logging data and build up the link between seismic attributes information and geologic and drilling information.The paper listed the applied cases of two oilfields and further proved application effects of the method.
引文
[1] BergeT B et al.Seismic inversion successfully predictsreservoir,porosity,and gas content inIbhubesiField,OrangeBasin,SouthAfrica.TheL eadingEdge,2002,4:338~348
    [2] 陈遵德.储层地震属性优化方法.北京:石油工业出版社,1998
    [3] 李庆忠.走向精确勘探的道路.北京:石油工业出版社,1994
    [4] 凌云研究组.基本地震属性在沉积环境解释中的应用研究.石油地球物理勘探,2003,38(6):642~653

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