多点地质统计学在储层相建模中的应用

Application of multiple-point geostatistics in reservoir facies modeling

刘超¹ 谢传礼¹ Yannick Tepinhi¹ 汪磊²

中国石油大学（北京）石油工程学院，北京 102249 中国石化胜利油田分公司河口采油厂，山东 东营 257051

College of Petroleum Engineering, China University of Petroleum, Beijing 102249, China Hekou Oil Production Plant, Shengli Oilfield Company, SINOPEC, Dongying 257051, China

针对渤南油田五六区沙三段4砂组2小层（Es₃4²）储层砂体形态弯曲、分叉、砂体规模与展布变化大的特点，利用多点地质统计学建模法对研究层位进行沉积微相模拟。在该区密集井网区精细地层研究的基础上，统计扇三角洲前缘水下分流河道与河口坝砂体长、宽定量数据，结合储层地质知识库的研究，确立不同微相的平面形态与空间组合特征，依此建立定量训练图像。在此基础上，利用SNESIM算法对研究区扇三角洲前缘进行沉积微相模拟。模拟结果表明，利用多点地质统计方法，能够很好地模拟出扇三角洲前缘水下分流河道与河口坝的空间形态特征，得到符合地质意义的沉积微相模型。研究区密集井网的定量研究和训练图像的建立，可以丰富地质知识库，为类似地区利用多点地质统计学进行储层相建模时训练图像的建立提供指导。

The reservoirs of 42 sand formation in Shahejie Group, Eocene of Bonan Oilfield （Es342） are front fan?鄄delta, which are of being meandering and forking with variable sand body distribution and scale. In this regard, the multi-point geostatistics simulation （MPS） is applied to model this spatial complicated microfacies. The first and core study is to get quantitative statistical data of underwater distributary channel and sand bar of front fan delta sediment in focused area on the base of stratigraphic study, combining 3D microfacies characteristics acquired from reservoir geological database, the quantitative training image is built. On the basis of established training image, SNESIM algorithm is applied to simulate microfacies property of studied area. The result shows that MPS can reconstruct paleo-microfacies sand body（underwater distributary channels and sand bar）reasonably in studied area and the micro-facies model from simulation is coherent to geological knowledge. The study on quantitative measurements of geological bodies in dense-well region and establishing of training image provide a guidance for other regions which have similar geological conditions.

10.6056/dkyqt201506017