基于嵌入式离散裂缝模拟的裂缝性油藏缝网分布反演

The fractured reservoir, which has changeable lithofacies and strong heterogeneity features, accounts for more than 28% of proven oil reserves. Aimed at detailed description of fracture network forms and distribution, fractured reservoir simulation and hierarchical fracture network distribution inversion research is carried out by using a combined method of computation simulation, inverse problem modeling and optimization, which is circle on the key scientific problems as complex simulation of fracture reservoir and uncertainty of fracture network distribution. The content is as follows:(1)embedded discrete fracture reservoir simulation;(2) macroscopic inversion prediction based on the level set theory in high permeability area of fracture network distribution;(3)precisely inversion prediction based on the principal component analysis theory of discrete fracture network distribution. With the implementation of this project, a practical method would be formed by utilizing dynamic data during reservoir development to inverse underground fracture network distribution, which could provide solutions and theory support for uncertainty problems of fracture network during fracture reservoir development and enhance oil recovery.

摘要：裂缝性油藏占探明石油储量的28%以上，具有岩相变化大、非均质性强等特点。围绕裂缝性油藏复杂渗流模拟、缝网分布不确定性的关键科学问题，以准确描述缝网形态和分布为目标，采用计算模拟、反问题建模与优化求解相结合的方法，开展裂缝性油藏数值模拟与分层次缝网分布反演研究，主要内容包括：（1）嵌入式离散裂缝油藏数值模拟方法；（2）基于水平集理论的缝网分布高渗区域宏观反演预测；（3）基于主成分分析理论的离散裂缝缝网分布精细反演预测。通过本项目的实施，将形成一套利用油藏开发动态数据反演地下缝网分布的实用方法，为裂缝性油藏开发的缝网不确定性问题提供解决思路与理论支持，提高油藏采收率。

油藏内裂缝的分布对油田开发效果影响很大，通过裂缝性油藏反演可以确定缝网的分布范围。常规的裂缝性油藏反演是使用油藏数值模拟软件并结合人为经验完成的，由于裂缝缝网分布具有极大的不确定性，油藏数值模拟软件难以满足需求，同时反演过程对人为经验依赖性大，不利于现场实际应用。本课题提出建立嵌入式离散裂缝网络模型，根据等效、水平集理论、数据分解压缩重构方法，并结合各种优化算法，对裂缝缝网进行精细反演。通过对微小裂缝的等效处理、非结构化网格剖分以及模拟有限差分法求解，建立的嵌入式离散裂缝网络模型切合实际的裂缝性油藏；对于高渗通道的离散特征，基于水平集理论，将离散的高渗通道进行转化，完成了离散高渗通道的反演预测；针对大规模裂缝缝网反演问题，采用数据分解压缩重构方法，对油藏参数降维，对比不同的优化算法，基本实现了对大规模裂缝缝网反演预测。在国内外发布期刊论文和会议论文28篇，22篇被SCI检索，2篇被EI检索。获批国家专利3项，参加国内外学术会议11次，培养研究生8名，完成了项目任务书的既定研究计划，研究成果能够为实际油田开发中的裂缝性油藏反演提供理论支撑。