天然气地球科学
高级检索
作者投稿 专家审稿 编辑办公

天然气地球科学 ›› 2020, Vol. 31 ›› Issue (6): 809–817.doi: 10.11764/j.issn.1672-1926.2019.12.004

• 天然气地质学 • 上一篇    下一篇

基于众数法聚类的多点地质统计学方法

麻平山1(),李少华1(),卢昌盛1,黄导武2,段冬平2,陆嫣2,丁芳2,黄鑫2   

  1. 1.长江大学地球科学学院, 湖北 武汉 430100
    2.中海油上海分公司研究院, 上海 200030
  • 收稿日期:2019-05-07 修回日期:2019-12-01 出版日期:2020-06-10 发布日期:2020-06-17
  • 通讯作者: 李少华 E-mail:2059200152@qq.com;534354156@qq.com
  • 作者简介:麻平山（1995-），男，山西朔州人，硕士研究生，主要从事地质统计学建模算法研究.E?mail:2059200152@qq.com.
  • 基金资助:
    国家科技重大专项(2016ZX05027-004);国家自然科学基金(41872129);湖北省自然科学基金创新群体项目(2016CFA024)

Multi-point geostatistical method based on mode method clustering

Ping-shan MA1(),Shao-hua LI1(),Chang-sheng LU1,Dao-wu HUANG2,Dong-ping DUAN2,Yan LU2,Fang DING2,Xin HUANG2   

  1. 1.College of Geosciences, Yangtze University, Wuhan 430100, China
    2.Shanghai Company Ltd. , CNOOC, Shanghai 200030, China
  • Received:2019-05-07 Revised:2019-12-01 Online:2020-06-10 Published:2020-06-17
  • Contact: Shao-hua LI E-mail:2059200152@qq.com;534354156@qq.com
  • Supported by:
    The National Science and Technology Major Project(2016ZX05027-004);The National Natural Science Foundation of China(41872129);The Natural Science Fund Innovation Group Project of Hubei Province of China(2016CFA024)

RichHTML

6

PDF (PC)

442

摘要：

多点地质统计学在建立复杂结构的地质模型方面应用较为广泛。基于样式的多点地质统计学方法SIMPAT在提取数据样式、计算数据样式相似度上效率存在不足。笔者提出了一种基于众数法聚类的多点地质统计学方法（CMMS）。基本思想是定义大小2类模板，通过大模板扫描训练图像提取数据样式，并使用小模板对大模板提取的数据样式进行降维。针对离散型变量，把大模板提取的数据样式中出现次数最多的相作为对应小模板的相类型。针对连续型变量，把数据样式中各网格的变量平均值作为对应小模板的值。相似的样式通过小模板降维可以聚为一类，这种降维方式对大模板提取数据样式包含的信息保真度较高。模拟时先比较数据事件与聚类样式代表，找出最相似的一类样式，然后在选取的样式类中寻找最匹配的数据样式。通过2次样式的比对，大大减少样式相似度计算的次数。通过实例的模拟效果、计算时间对比、统计特征再现等测试表明，CMMS算法在样式对比上的计算效率比SIMPAT提高30~40倍，CMMS的模拟结果能够很好地再现训练图像中的地质模式及统计特征，并且该方法能够直接应用到连续型变量的模拟。

关键词: 多点地质统计学, 聚类, 计算效率, SIMPAT, 数据样式

Abstract:

Multi-point geostatistics is widely used in the establishment of geological models of complex structures. The pattern-based multi-point geostatistical method SIMPAT has insufficient efficiency in extracting data patterns and calculating their similarities. This paper proposes a multi-point geostatistical method based on the cluster method (CMMS). The basic idea is to define two types of templates, scanning training image through the large template to extract the data pattern, and use the small template to reduce the dimension of the data pattern. For discrete variables, the facies with the largest number of occurrences in the data pattern extracted from the large template is taken as the facies type of the corresponding small template. For continuous variables, the average values of variables of each grid in the data pattern are taken as the values of corresponding small templates. Similar patterns can be grouped into one group by dimensionality reduction through small templates, which has a high fidelity of information contained in data patterns extracted from large templates. In the simulation, data events are compared with clustering pattern representation to find the most similar pattern, and then the most matching data pattern is found in the selected pattern class. By two times comparison, the number of similarity calculation is greatly reduced. The results of simulation, calculation time comparison and statistical feature reproduction show that the computational efficiency of CMMS algorithm is 30-40 times higher than that of SIMPAT. The simulation results of CMMS can well reproduce the geological model and statistical features in training images, and this method can be directly applied to continuous variables simulation.

Key words: Multi-point geostatistics, Clustering, Computation efficiency, SIMPAT, Data pattern

中图分类号: 

  • TE19

图1

Fluvsim 69×69×39训练图像"

图2

粗网格内众数法聚类原理"

图3

模拟流程"

图4

训练图像及模拟结果"

表1

计算20个模拟实现平均耗时"

训练图像计算时间/s
SIMPATCMMS
二维1584.3
三维1 53544

图5

相似度统计直方图"

表2

参数敏感性测试"

网格重数样式大小

图6

不同类型比例统计直方图"

图7

变差函数对比"

1 肖斌,赵鹏大,侯景儒.地质统计学新进展[J].地球科学进展,2000，15(3):293-296.
XIAO B,ZHAO P D,HOU J R. New development of geostatistics[J].Advances in Earth Science,2000，15(3):293-296.
2 尹艳树,张昌民,李玖勇,等.多点地质统计学研究进展与展望[J].古地理学报,2011,13(2):245-252.
YIN Y S，ZHANG C M，LI J Y，et al. Progress and prospect of multiple-point geostatistics[J]. Journal of Palaeogeography,2011,13(2):245-252.
3 吴胜和,李文克.多点地质统计学——理论、应用与展望[J].古地理学报,2005，7(1):137-144.
WU S H,LI W K.Multiple-point geostatistics: Theory, application and perspective[J]. Journal of Palaeogeography,2005，7(1):137-144.
4 贾爱林.中国储层地质模型20年[J].石油学报,2011,32(1):181-188.
JIA A L.Research achievements on reservoir geological modeling of China in the past two decades[J]. Acta Petrolei Sinica,2011,32(1):181-188.
5 GUARDIANO F,SRIVASTAVE M. Multivariate Geostatistics: Beyond Bivariate Moments[M]∥ Dordrecht: Kluwer Academic Publication, 1993：133-144.
6 STREBELLE S B. Sequential Simulation Drawing Structures From Training Images[D].Palo Alto:Stanford University,2000: 40-50.
7 STREBELLE S B. Conditional simulation of complex geological structures using multiple-point statistics[J]. Mathematical Geology, 2002, 34(1): 1-21.
8 GUVEN B A. Sequential Simulation with Patterns[D]. Palo Alto:Stanford University, 2005:10-15.
9 ZHANG T F, SWITZER P, JOURNEL A. Filter-based classification of training image patterns for spatial simulation[J]. Mathematical Geology,2006,38(1): 63-80.
10 MARIETHOZ G, RENARD P, STRAUBHAAR J. The direct sampling method to perform multiple‐point geostatistical simulations[J]. Water Resources Research, 2010, 46(11):1-14.
11 REZAEE H, MARIETHOZ G, KONESHLOD M, et al. Multiple-point geostatistical simulation using the bunch-pasting direct sampling method[J]. Computers & Geosciences,2013,54: 293-308.
12 HONARKHAH M, CAERS J. Stochastic simulation of patterns using distance-based pattern modeling[J]. Mathematical Geosciences,2010, 42(5): 487-517.
13 HANSEN, MEJER T, MOSEGAARD K, et al. Multiple point statistical simulation using uncertain (soft) conditional data[J].Computers & Geosciences,2018, 114: 1-10.
14 GARDET C, RAVALEC M, GLOAGUEN E. Pattern-based conditional simulation with a raster path: A few techniques to make it more efficient[J].Stochastic Environmental Research and Risk Assessment,2016,30(2): 429-446.
[1] 吴浩, 刘锐娥, 纪友亮, 张春林, 周勇, 张云钊. 典型致密砂岩气储层孔隙结构分类及其意义——以鄂尔多斯盆地盒8段为例[J]. 天然气地球科学, 2016, 27(5): 835-843.
[2] 夏明军, 曾大乾, 邓瑞健, 姜贻伟, 毕建霞, 靳秀菊, 谭国华, 苗菁. 普光气田长兴组台地边缘礁、滩沉积相及储层特征[J]. 天然气地球科学, 2009, 20(4): 549-556,562.
[3] 尹艳树,吴胜和. 储层随机建模研究进展[J]. 天然气地球科学, 2006, 17(2): 210-216.
[4] 尹太举;张昌民;王寿平;. 濮53块流动单元分类方法研究[J]. 天然气地球科学, 2005, 16(3): 298-301.
[5] 刘刚. 灰色聚类分析在圈闭评价中的应用[J]. 天然气地球科学, 2001, 12(6): 18-22.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!