【摘要】 多点地质统计学综合了基于象元方法以及基于目标方法两者的优点,对于河流相等具有复杂地质形态的储层精确建模具有较强的优势。在对传统建模方法综合分析的基础上,介绍了多点地质统计学的基本理论及SNESIM算法,并应用该技术对大牛地气田某开发井区的辫状分流河道相进行了实际建模。研究结果表明,在河流相储层建模中,该方法比传统的建模方法更具优越性。最后,进一步综合讨论了多点地质统计学目前面临的主要问题(包括训练图像、目标体连续性、数据样板选择、综合地震信息等方面)的改进方法。更多还原

【Abstract】 Multiple-point geostatistics is a promising discipline in reservoir stochastic modeling. This approach combines the advantages of two methods: pixel-based two-point simulation and object-based simulation, which is able to make a more exact reservoir modeling for the reservoir with complex variability, especially the fluvial reservoir. With the analysis of the traditional modeling methods, the paper presents the principle of multiple-point geostatistics and SNESIM algorithms, and simulates the sand distribution of a braided distributary channel reservoir in one development block, Daniudi gas field. The result indicates that this approach is better than the traditional methods for the fluvial reservoir modeling. Finally, the paper discusses some main problems, including the training images, object continuity, data template, and integration of seismic information.更多还原