Two-Phase Fluid Leakage through Faults Using a Multi-Scale Analytical-Numerical Modeling Approach

Friday, 19 December 2014: 10:50 AM
Mary Kang, Princeton University, Princeton, NJ, United States; Stanford University, Los Altos Hills, CA, United States, Jan M Nordbotten, University of Bergen, Bergen, Norway, Florian Doster, Heriot-Watt University, Edinburgh, United Kingdom and Michael A Celia, Princeton University, Civil and Environmental Engineering, Princeton, NJ, United States
Fluid flow through faults must be considered in many applications including geologic storage of carbon dioxide (CO2), deep storage of hazardous waste, groundwater contamination, and petroleum engineering. In the case of COstorage, the presence of faults is of concern, because they can act as leakage pathways. Therefore, modeling tools that can accurately and efficiently quantify fluid leakage through faults in basin-scale models are necessary. In basin-scale models, the flow around and through faults is a local-scale process and this local-scale variation is important when determining leakage rates. We present a multi-scale modeling approach based on embedding local-scale analytical solutions within basin-scale numerical models.

At the local scale, steady-state analytical solutions that represent fluid flow in the vicinity of leaky faults, including any vertical flow effects, are derived. Using both numerical simulations and analytical solutions, an empirical model representing fault properties, permeabilities and widths, is also developed. The combination of this empirical fault model and the analytical solutions captures the local-scale effects of leakage through faults.

The local-scale model is used within a multi-scale modeling framework to determine the flow in and around faults and the associated local-scale pressure and saturation corrections that are applied to the coarse model. Here, a fault is viewed as a 2-D surface on one side of a coarse-scale grid block. The corrections relate local-scale pressure and saturation at the fault to coarse-scale pressures and saturations in numerical grid blocks. The corrections are used to determine the vertical and lateral flow in the fault and horizontal flows perpendicular and parallel to the fault in the grid block. At every coarse-scale time step, the local-scale fault model is implemented using the coarse-scale information from the previous time step. The resulting leakage rates and pressure and saturation corrections are updated in the coarse-scale model by modifying the coefficients of the coarse model.

Simulations using this multi-scale model will be shown for (1) a fault that extends through a single grid block, and (2) a fault that extends through multiple grid blocks.