摘要
针对常规地质建模方法在辫状河致密砂岩气藏适应性差的问题,文中以鄂尔多斯盆地苏里格气田为研究对象,提出以地质知识库指导三维训练图像建立的多点地质统计学建模方法。首次采用Google Earth建立现代露头解剖地质知识库,在现有井网条件下,结合沉积特征提出3种井钻遇心滩坝位置的预测方法,建立了研究区定量化地质知识库,定量表征了心滩坝的长、宽、厚等几何参数。参考心滩坝构型模式,建立了定量化的三维心滩坝训练图像,利用多点地质统计学Snesim算法进行地质建模,并与传统序贯指示模拟方法对比分析。结果表明:序贯指示模拟方法仅能在平面上表征心滩坝特征,垂向上不符合地质认识,而多点地质统计学方法建立的模型,能够定量化地表征心滩坝的几何形态和空间结构。
Considering the poor applicability of conventional geological modeling to tight sand gas reservoir of braided river facies, a modeling method of multiple-point geostatistics for 3 D training image establishment based on geological knowledge was put forward taking Sulige gas field in Ordos Basin as the study object. It is the first time to use Google Earth to establish a modern outcrop anatomical geological knowledge database. Based on the existing well pattern and combined with the depositional characteristics,three forecasting methods for drilling location of diara dam are put forward. Quantitative geological knowledge base for study area is established to quantitatively represent the geometric parameters such as length, width and thickness of diara dam. Quantitative 3 D diara dam training image is built with reference diara dam model, geological modeling using the multiple-point geostatistics Snesim algorithm is compared with traditional sequential instruction simulation method. The results show that the sequential instruction simulation method can only represent the characteristics of the diara dam in the plane and it is not consistent with the geological knowledge while the multiple-point geostatistics method can be used to model the geometric and spatial structure of diara dam.
出处
《断块油气田》
CAS
北大核心
2018年第1期57-61,共5页
Fault-Block Oil & Gas Field
基金
国家科技重大专项专题“深层礁滩相储层气水两相渗流特征研究”(2016ZX05017-001-HZ02)
关键词
多点地质统计学
地质知识库
地质建模
心滩坝
辨状河
multiple-point geostatistics
geological knowledge database
geological modeling
diara dam
braided river
