基于三维训练图像的多点地质统计学岩相建模  

Application of multiple-point geostatistics lithofacies modeling base on 3D training image

作　　者：刘跃杰[1,2] 李继红[1,2] 陈卓[1,2] 杜康[3] 

机构地区：[1]西北大学地质系,陕西西安710069 [2]西北大学大陆动力学国家重点实验室,陕西西安710069 [3]中国石化西北油田分公司采油二厂,新疆轮台841600

出　　处：《石油化工应用》2015年第9期94-100,共7页Petrochemical Industry Application

基　　金：中国石油天然气集团公司重大攻关项目"鄂尔多斯盆地致密油勘探开发关键技术研究";项目编号:2011E-13

摘　　要：针对多点地质统计学算法在三维区块模拟中获取合适的训练图像较为困难的问题,尝试用多点地质统计学方法建立了马岭油田长821小层的三维岩石相模型。首先介绍了多点地质统计学的基本原理和方法;然后结合已有的地质认识和相关的文献等资料建立了各岩相的空间接触关系,并对砂岩含量属性图阀值化建立起了研究区长821小层的训练图像;最后以井上解释的岩相数据为硬数据,各岩相比例统计成果为软数据,利用多点地质统计学中的Snesim算法实现了对长821小层的岩石相建模,并通过一系列模型检验方法对训练图像的维数、各岩相比例进行了对比分析;研究结果表明:在各岩相比例合适的前提下,采用阀值化方法获取的三维训练图像使多点地质统计学能更好的适应本区的岩石相建模,再现不同岩相砂体的复杂空间分布特征,为马岭油田有效储集砂体的预测提供了新的解决手段,对马岭油田的进一步开发具有重要指导意义。Provided that get proper training in a simulated three-dimensional image of the block is a more difficulty for multiple-point geostatistics algorithms problem, in this paper multiple-points geostatisties is introduced to the 3D lithofacies model of the Chang82~ sublay- er in Maling oilfield.Firstly,basic concept and method of multiple-point geostatistics are in- troduced in the paper.Secondly,the space contact relationships of different lithofacies is es- tablished through comprehensive analysis of all kinds of materials including geological knowledge and related literature material,and then the lithofaeies training image of the Chang82~ sublayer is made by the methods of threshold to the sand-body attribute datum. Finally,with the hard data of well point lithofacies interpretation and soft data of the various lithofacies proportion,three-dimensional lithofacies is created with the method of multiple- point geostatistics of Snesim algorithms,and comparative analysis to training image dimension and the proper of the various lithofacies proportion.In the proper proportion of each lithofa- cies premise, use threshold method to obtain three-dimensional images to make multiple- point geostatistics can better adapt to the area＇s lithofacies modeling and reproduce the com- plex spatial structure and configuration of the various lithofacies,which provides a new mea- surement to predict the distribution of effective sand-body and has a guiding significance for further development at Maling oilfield.

关 键 词：多点地质统计学 训练图像 Snesim算法 岩相模型 

分 类 号：P628.2[天文地球—地质矿产勘探] TP391.41[天文地球—地质学]