分区域多点统计随机地质建模方法——以柴达木盆地辫状河三角洲沉积储集层为例  

Regionalized multiple-point stochastic geological modeling: A case from braided delta sedimentary reservoirs in Qaidam Basin, NW China

作　　者：陈更新[1,2] 赵凡[2] 王建功[2] 郑红军[2] 严耀祖[2] 王爱萍[2] 李积永 胡云鹏[4] 

机构地区：[1]中国地质大学(北京)能源学院 [2]中国石油勘探开发研究院西北分院 [3]中国石油青海油田勘探开发研究院 [4]中国石油勘探开发研究院

出　　处：《石油勘探与开发》2015年第5期638-645,共8页Petroleum Exploration and Development

基　　金：国家科技重大专项(2011ZX05007-006);中国石油重大科技专项(2011E-0308)

摘　　要：通过对密井网区精细地质研究,结合沉积物理模拟实验,对柴达木盆地昆北油田辫状河三角洲沉积模式开展研究,获得沉积微相单元的统计特征参数和平面配置关系,建立辫状河三角洲不同沉积区域的训练图像,最终建立辫状河三角洲储集层地质模型。采用角度旋转、比例缩放、邻区序贯条件模拟方法,分区分训练图像建立精细的辫状河三角洲沉积微相模型。分区域多点统计随机建模间接放宽多点统计随机建模平稳性要求,并改善了河道连续性,提高了训练图像的可移植性。在柴达木盆地昆北油田油藏描述应用中,模型与生产实际吻合较好,取得较好应用效果,证明该方法的可靠性和可行性,探索了采用随机模拟方法准确刻画辫状河三角洲沉积储集层的新途径。Research on the depositional model of a braided delta in Kunbei oilfield, Qaidam Basin, was conducted with a fine geologic study in a dense well network area by combining sedimentary physical simulation to obtain statistical characteristic parameters and allocation relations of plane sedimentary facies units. On this basis, training images in different sedimentary regions of the braided delta were developed for reservoir modeling. In each sub-region, sequential conditional simulation in adjacent regions used different training images with angle revolving and scaling. This simulation succeeded in finely modeling the sedimentary microfacies of the braided delta. This method relaxed the requirement for a stationary training image in multipoint simulation and improved channel continuity as well as portability of the training image. In the fine reservoir description of the Kunbei oilfield, the model results were consistent with practical production and have the potential to guide effective water-flooding development. Effective application demonstrated the feasibility and practicability of the method. This paper presents a new method for accurately characterizing the braided delta using stochastic simulation.

关 键 词：分区域多点统计 随机地质建模 训练图像 沉积微相 辫状河三角洲 柴达木盆地 

分 类 号：TE122[石油与天然气工程—油气勘探]