摘要：

低渗油藏微观孔隙结构高度复杂且差异显著，孔隙结构参数与渗透率之间的强非线性耦合进一步加剧了驱替方式优选的难度。为构建有效的孔隙结构分类体系并优选高效适配的驱替方式，本文采用数据挖掘方法对鄂尔多斯盆地延长组某区块470块低渗致密砂岩岩心的压汞数据开展系统分析，将孔隙结构划分为大尺度弱非均质型、大尺度强非均质型、小尺度弱非均质型和小尺度强非均质型4类；在此基础上，选取各类典型孔径分布建立毛管束模型开展驱替模拟，以阈值压力为判据划分气驱与水驱的优势区间，定量评价两种驱替方式的效果。研究表明，气驱与水驱的相对优势主要受孔隙尺度、非均质性与注入压力的耦合影响。大尺度弱非均质条件下，低压气驱占优，随着注入压力升高水驱因前缘稳定与动用孔隙增多逐步占优，阈值压力随注入体积增大而增大；大尺度强非均质性显著加剧气窜并压缩气驱占优的压力区间。对于小尺度孔隙，水驱需高压方可实现有效驱替并在较高压差下才会体现出优势，此时工程可实施性成为关键约束；小尺度强非均质虽进一步放大气窜劣势但同时也削弱水驱增压收益，且阈值压力仍偏高，在可实施压差受限时可优先考虑低压气驱等方案。本文建立的孔隙结构分类与驱替效果评价方法，可为低渗油藏在可注入能力约束下开展驱替方式优选与提高采收率策略制定提供依据。

关键词: 孔隙结构分类, 低渗油藏, 数据挖掘, 毛管束模型, 驱替效果评价

Abstract:

Low-permeability reservoirs exhibit highly complex and strongly variable pore systems. The pronounced nonlinear coupling between pore-structure attributes and permeability further complicates displacement process screening. To establish a pore-structure classification framework and identify suitable displacement schemes, this study applies data-mining techniques to mercury intrusion capillary pressure data from 470 low-permeability tight sandstone cores collected from a target block of the Yanchang Formation in the Ordos Basin. The pore systems are classified into four types: large pore-size with weak heterogeneity, large pore-size with strong heterogeneity, small pore-size with weak heterogeneity, and small pore-size with strong heterogeneity. Representative pore-size distributions for each type are then used to build capillary-bundle models for displacement simulations. A threshold pressure differential is adopted as the criterion to delineate the favorable operating windows for gas flooding and water flooding, and the performance of the two processes is quantitatively evaluated. The results show that the relative superiority of gas flooding versus water flooding is mainly controlled by the coupled effects of pore size, heterogeneity, and injection pressure differential. For large pore-size systems with weak heterogeneity, gas flooding is preferred at low pressure differentials; as the pressure differential increases, water flooding gradually becomes dominant due to improved front stability and a larger swept pore volume, and the threshold pressure differential increases with increasing pore volumes injected. In contrast, strong heterogeneity in large pore-size systems significantly intensifies gas channeling and early breakthrough, narrowing the pressure range in which gas flooding is advantageous. For small pore-size systems, effective water flooding requires sufficiently high pressure differentials and becomes advantageous only under relatively large pressure gradients, making operational feasibility a key constraint. In small pore-size systems with strong heterogeneity, heterogeneity further aggravates gas channeling while reducing the incremental benefit of increasing waterflood pressure; the threshold pressure differential remains high. When allowable pressure differentials are limited, low-pressure gas flooding may be prioritized as a practical alternative. Overall, the proposed workflow integrating pore-structure classification and displacement-performance evaluation provides a quantitative basis for displacement process screening and enhanced oil recovery strategy design in low-permeability reservoirs under injectivity constraints.

Key words: pore-structure classification, low permeability reservoirs, data mining, capillary bundle model, displacement-performance evaluation

中图分类号: 

• TE311

• TE348