摘要
随着水驱开发老油田进入深度精细开发阶段,地下油水分布动态复杂化对油藏管理提出更高要求。系统综述流线模拟技术在油藏开发中的研究进展,梳理其从二维追踪算法到三维多相流模型、再到智能化阶段的发展历程,阐明了通过降维(三维渗流场分解为一维流线)求解提升计算效率的算法原理。将流线模拟技术分为流线模拟和流线表征两个部分,前者通过求解压力场与饱和度场,揭示流体在多孔介质中的渗流规律;后者依托速度场重构、流线路径追踪与多物理场参数传递,完成流动轨迹可视化建模。系统总结了流线模拟技术在辅助历史拟合与模型优化、油藏模型粗化与计算优化、复杂驱替机理与精细表征、油藏管理与优化、注水与井网优化、油藏动态分析与不确定性分析等方面的应用成果。当前流线模拟技术多物理场耦合建模、跨尺度计算效率及数据依赖性等方面存在技术瓶颈,亟需通过机器学习、自适应网格优化与高性能计算深度融合,为能源可持续发展提供支持。
As water-flooded mature oilfields enter the stage of deep and refined development,the complex dynamics of subsurface oil-water distribution impose higher demands on reservoir management.This paper systematically reviews the research progress of streamline simulation technology for reservoir development,outlining its developmental trajectory from two-dimensional tracing algorithms to three-dimensional multiphase flow models,and further to the intelligent stage.The algorithmic principle of enhancing computational efficiency through dimensionality reduction(decomposing three-dimensional flow fields into one-dimensional streamlines)is elucidated.The streamline technology is categorized into two components:streamline simulation and streamline characterization.The former reveals fluid flow patterns in porous media by solving pressure and saturation fields,while the latter accomplishes visual modeling of flow trajectories through velocity field reconstruction,streamline path tracing,and multiphysical parameter transfer.The study comprehensively summarizes applications of streamline simulation in aiding history matching and model optimization,reservoir model upscaling and computational acceleration,characterization of complex displacement mechanisms,reservoir management and optimization,water injection and well pattern optimization,as well as dynamic analysis and uncertainty quantification.However,current streamline simulation technology faces challenges in multi-physical field coupling modeling,across-scale computational efficiency,and data dependency.Breakthroughs in machine learning,adaptive mesh optimization,and high-performance computing integration are urgently needed to support sustainable energy development.
作者
朱冰倩
钱其豪
张立侠
黎镇东
ZHU Bingqian;QIAN Qihao;ZHANG Lixia;LI Zhendong(Research Institute of Petroleum Exploration and Development,PetroChina,Beijing 100083,China)
出处
《石油钻采工艺》
北大核心
2025年第2期131-143,共13页
Oil Drilling & Production Technology
基金
中国石油天然气股份有限公司科技项目“老油田‘压舱石工程’关键技术研究与示范”(编号:2023YQX10201)
中国石油天然气集团有限公司科技项目“提高原油采收率新方法与新技术研究”(编号:2023ZZ0403)。
关键词
流线模拟
流线追踪
油藏开发
多物理场耦合
机器学习
streamline simulation
streamline tracing
reservoir development
multiphysics field coupling
machine learning
