基于Minkowski泛函和K-means聚类算法的岩石类型划分  

Rock-type Classification Based on Minkowski Functionals and K-means Cluster Analysis

作　　者：杨博[1] 刘钰洋[1] 潘懋[1] 

机构地区：[1]北京大学地球与空间科学学院石油与天然气研究中心,北京100871

出　　处：《科学技术与工程》2017年第22期63-67,共5页Science Technology and Engineering

基　　金：国家科技重大专项(2016ZX05010-001;2011ZX05020-008)资助

摘　　要：数字岩心技术的发展和计算科学、物理学、统计学、油藏工程等多学科的交叉融合,为基于数字岩心图像的岩石类型的精细划分提供了一种新的方法。基于micro-CT物理法成像处理后获得的砂岩数字岩心图像,运用闵可夫斯基泛函作为基于数字岩心图像的岩石物性特征的综合表征参数;并结合K-means聚类算法,可以达到基于数字岩心图像的岩石类型的精细划分和获得基于精细岩石类型层次系统的目的。对于后续基于精细岩心层次系统岩心物性参数的模拟和粗化具有十分重要的意义。A new approach is provided in elaborate rock-type classification based on digital core images with the development and amalgamation of computation science,physics,statistics and reservoir engineering.Based on the digital sandstone core images which gets from the micro-CT facility,Minkowski functionals is performed as the integration character of rock physical properties along with the K-means algorithm to achieve the goal of elaborate rocktype classification and getting the rock sample＇s layer system which plays an important part of rock physical properties simulation and up scaling based on the layer system.

关 键 词：砂岩数字岩心 岩石类型划分 闵可夫斯基泛函 K-means聚类分析 

分 类 号：TE343[石油与天然气工程—油气田开发工程]