首页|基于矢量信息的多点地质统计学算法

基于矢量信息的多点地质统计学算法

扫码查看
在Snesim算法的基础上,提出基于地质矢量信息的多点地质统计学算法(VMPS).以冲积扇为例,研究冲积扇地质矢量坐标系统,并在训练图像中融入地质矢量信息,形成基于矢量信息的训练图像.同时,对训练图像扫描机制进行改进,利用矢量信息统计数据事件和区域相比例,建立基于地质矢量信息的搜索树.在模拟过程中根据当前模拟区块的矢量信息,提取与之关系最密切的数据事件计算重复数,得到适应当前模拟区域的条件概率分布,从而满足局部平稳的条件,再现不同沉积相带不同地质特征.通过概念模型和实际储层模拟检验表明:VMPS的模拟效果明显优于Snesim,能适应“非平稳性”突出条件下的地质建模.
A vector information based multiple-point geostatistic method
Take Snesim as algorithm foundation,a new multi-point geostatistics algorithm VMPS was presented on the basis of geological vector information.Taking alluvial fan as an example,the alluvial fan geological vector coordinate system was constructed,the geological vector information was integrated in training images,and the vector information based training images was formed.Then,the scanning mechanism for training images was improved,and the search tree was established based on the geological vector information with the vector information statistics events and regional conditional proportion between provides areas.In the simulation process,based on the vector information of the current simulation area,the data event was extracted to calculate the repeat numbers,so as to get the conditional probability distribution which fulfills the condition of local stability and establishes models that can reproduce different characteristics of different area.The ideal models and actual reservoir simulation test show that the simulation effect of VMPS is superior to that of Snesim,as VMPS can be applied to the reservoir modeling under the "non-stationary" conditions.

multi-point geostatics algonthnstochastic reservoir modelingvector informationalluvial faninformation integrating

冯文杰、吴胜和、印森林、石书缘、刘俊玲

展开 >

中国石油大学(北京)地球科学学院,北京,102249

油气资源与探测国家重点实验室(中国石油大学),北京,102249

中国石油勘探开发研究院,北京,100083

多点地质统计学算法 储层随机建模 矢量信息 冲积扇 多信息整合

国土资源部油气重大专项国家自然科学基金

2011ZX05009-00341372116

2014

中南大学学报(自然科学版)
中南大学

中南大学学报(自然科学版)

CSTPCDCSCD北大核心EI
影响因子:0.938
ISSN:1672-7207
年,卷(期):2014.45(4)
  • 8
  • 10