运用多点地质统计学确定水下分流河道宽度及钻遇概率

向传刚

Xiang Chuangang

中国石油大庆油田有限责任公司勘探开发研究院，黑龙江 大庆 163712

Research Institute of Exploration and Development, Daqing Oilfield Company Ltd., PetroChina, Daqing 163712, China

针对传统地质统计学难以定量描述空间变量的问题，文中运用多点地质统计学方法，分别以40，80，120，160，200 m制作训练图像，对某密井网区水下分流河道相进行了随机模拟，并对基础井网设置虚拟更新井，分不同方向统计虚拟对子井钻遇河道的概率，然后与实际对子井钻遇情况进行对比分析。结果表明，水下分流河道的宽度为120 m时，100次随机模拟结果的虚拟对子井钻遇河道的情况，与实际情况接近，可以确定其为研究区内水下分流河道的常见宽度。分析结果为该区储层构型精细研究奠定了基础。

In order to overcome the shortcomings of traditional geostatistics which is difficult to quantitatively describe the space variable, the training images were produced according to 40 m, 80 m, 120 m, 160 m and 200 m and by using the method of multiple-point geostatistics, and then the stochastic simulation was made on the underwater distributary channels in one density well area. Some virtual updating wells were set for every basic well and the probability of virtual paired wells drilling the underwater distributary channels was predicted based on the different directions, which were compared with the probability of actual paired wells drilling the underwater distributary channels. The results show that when the underwater distributary channel width is 120 m, the probability of virtual paired wells is close to that of actual paired wells based on 100 times stochastic simulation, which can be determined as the common width of underwater distributary channel in study area. Research results laid a solid foundation for studying reservoir architecture.

10.6056/dkyqt201502006