Long after CO2 is injected into a brine aquifer, most reservoir-scale fluid dynamic simulations predict large fractions of the original plume will become immobilized via capillary trapping and dispersed throughout the formation. We begin our analysis with a reservoir in this state and consider the effects caused by a depressurization of the domain (e.g. from a nearby production well or newly formed fracture between neighboring reservoirs) and predict the fraction of CO2 that will be remobilized as a result. We then model the dynamics of this remobilized CO2 in two distinct steps: (1) vertical rise within the reservoir, followed by (2) spreading of mobile CO2 into the far-field of the domain and justify this approach from a scaling analysis of the governing equations. We show that a model of relative permeability that takes account of insights from percolation theory near the minimum CO2 saturation leads to much more rapid rise and subsequent radial spreading of remobilized CO2 than a traditional empirical correlation such as the Brooks-Corey model. Furthermore, we find that over a broad range of remobilized CO2 mass fraction and Bond number, the radial extent of the mobile plume does not exceed a factor of 1.8 times the radius of the original immobilized CO2 region.

在将CO2注入盐水层之后很长时间，大多数油藏规模的流体动力学模拟预测，大部分原始羽流将通过毛细管捕集而被固定并分散在整个地层中。我们从处于这种状态的储层开始分析，并考虑域降压（例如，来自附近的生产井或相邻储层之间新形成的裂缝）引起的影响，并预测将要运出的CO2比例。然后，我们通过两个不同的步骤来模拟这种迁移的CO2的动力学过程：（1）储层内的垂直上升，然后（2）将移动的CO2扩散到该域的远场中，并通过对油田的规模分析来证明这种方法的合理性。控制方程。我们表明，相对渗透率的模型考虑了渗流理论在最小CO2饱和度附近的见解，比传统的经验相关性（如Brooks-Corey模型）导致更快的上升速度和随后径向移动的CO2扩散。此外，我们发现在广泛的固定CO2质量分数和键数范围内，活动羽流的径向范围不超过原始固定CO2区域半径的1.8倍。