致密气
磁导率
材料科学
石油工程
化学
地质学
水力压裂
膜
生物化学
作者
Lele Wang,Bing Wei,Junyu You,Wanfen Pu,Jinyu Tang,Jun Lu
出处
期刊:Energy
[Elsevier BV]
日期:2022-10-11
卷期号:263: 125696-125696
被引量:64
标识
DOI:10.1016/j.energy.2022.125696
摘要
This work provides a comprehensive workflow with practical guidelines for integrating fracture-geometry, rock stress-sensitivity, and gas molecular diffusion to evaluate the performance of a tight reservoir horizontal well induced by gas Huff–n–Puff (HnP). The laboratory measurements and field results were integrated in numerical simulation. The stress-sensitivity led the permeabilities of the matrix and fractures to significantly decrease during primary depletion and HnP stages, thereby impairing well productivity. Rigorous characterizations of fracture-geometry and rock stress-sensitivity are crucial for history-matching and oil recovery simulation. During produced gas HnP, the cumulative oil production was reduced by 4.6% original oil in place (OOIP) because of the stress-dependent permeability. The hysteresis in fracture-permeability noticeably decreased the transport of mass and pressure but hardly affected the short-term production owing to the rapid closure of fractures during puff stage. The final oil recoveries induced by gas diffusion were 0.4%, 0.04% and 0.35% OOIP by CO 2 , N 2 and produced gas HnP, respectively. The produced gas HnP process produced the highest efficiency of enhanced oil recovery than CO 2 and N 2 owing to its significant pressurizing effect and lower gas-oil ratio. Minimum-miscibility-pressure should not be the primary consideration for gas HnP in tight reservoirs. • A workflow with practical guidelines was established to investigate well performance. • Fracture-geometry was retrieved using field-measured microseismic data. • Rock stress-sensitivity was incorporated throughout the well production life. • Hysteresis significantly impacted the distribution of pressure and fluids. • Gas diffusion improved gas utilization and thus promoted oil production.
科研通智能强力驱动
Strongly Powered by AbleSci AI