致密储层多尺度一体化表征技术研究

The tight oil is an important unconventional oil and gas resource as well as coal bed methane and shale oil and gas. Tight reservoir generally has a widespread distribution, its source and reservoir rock are immediate vicinity. The traps are not obvious. The oil saturation are heterogeneity, and Low single well production. The reservoir forming conditions and laws have obviously differences among different sedimentary basins. Aim at the Gaotaizi layer of Qingshankou formation in Songliao basin has the characteristic with strong heterogeneity. On the basis of confirming the sedimentary microfacies, basic types of reservoir and Micro-Analysis in the target area. We build reservoir scale, rock scale, micron scale, sub-micron scale and nanoscale three-dimensional database of five different scales. Through the research and development of platform, it is possible to make an effective integration between different scales of macroscopic geological parameters and parameters of microscopic pore structure. And it can provide a scientific basis for high quality reservoir prediction and tight oil resources evaluation, if the multi-scale integrated study of tight reservoir come true. It also has the very important realistic and strategic significance for tight reservoir exploration and development.

致密油同煤层气、页岩油气等一样是十分重要的非常规油气资源，致密油藏一般具有大面积分布、源储紧邻、圈闭不明显、含油饱和度非均质、单井产量低等特点，不同沉积盆地的致密油成藏条件和规律差异性明显。针对松辽盆地齐家地区青山口组高台子油层致密砂岩储层的非均质性较强的特点，在靶区沉积微相、储层基本类型及微观分析测试等成果的基础上，通过建立油藏尺度、岩石尺度、微米尺度、亚微米尺度和纳米尺度5个不同尺度的三维数据库，通过平台算法研发，将不同尺度的宏观地质参数与微观孔隙结构参数有效整合，实现致密储层多尺度一体化研究，为优质储层预测以及致密油资源评价提供科学依据，对致密储层的勘探开发具有十分重要的现实和战略意义。

松辽盆地北部齐家地区高三、高四油层组致密油资源丰富，是大庆油田重要的接替资源。齐家地区致密砂岩储层虽然整体呈大面积、层状、连续或准连续分布，但其存在储层非均质性强、致密油富集规律不明确。针对靶区致密砂岩储层的复杂性，本次研究综合利用了铸体薄片技术、扫描电镜技术、数字岩心分析技术、压汞技术等油气储层分析方法，通过建立不同尺度的三维数据库与平台算法研发，将不同尺度的宏观地质参数与微观孔隙结构参数有效整合，实现致密储层多尺度一体化研究。取得了如下结果与认识：（1）本研究在岩石和物性特征分析的基础上，通过多种资料综合分析识别出靶区6种沉积微相类型，并通过取心井的测井曲线识别特征要素，建立测井微相模式；（2）利用测井相模式，结合靶区沉积背景、微相平面组合、亚相理论模式等技术完成靶区41个小层的平面沉积微相识别并将其分为三角洲前缘亚相及前三角洲亚相，并且构建了宏观沉积微相与微观孔隙结构之间的联系，开展了综合分类研究；（3）基于机器学习的思想，通过支持向量机知类别的测井曲线数值进行学习并形成支持向量机模型，通过形成的支持向量机模型来对未测样点的微观孔喉分级模式进行预测，最终构建了一套致密储层微观-宏观一体化的表征方法。