Fluid dynamic modeling of multiphase flow in heterogeneous porous media with matrix, fracture, and skin

Underground contaminants such as non-aqueous phase liquids (NAPLs) or volatile organic compounds (VOCs) threaten public safety and health. The presence of fractures or regions of high permeability significantly limits the recovery due to the contrast in the capillary pressures of the low- and high-permeability regions. We previously proposed a methodology to improve the recovery of liquids upon gas injection process from a porous system with a percolating network of fractures. This methodology uses a high capillary pressure skin at the production well and controls the injection pressure below a threshold value for the breakthrough. In this paper, a mixed finite element method is employed to model the gas-water drainage process in a porous medium containing, matrix, fracture, and skin. An additional mixed-layer is also used to account for the mixing of the solid particles at the boundaries between the matrix, fracture, and skin. The mathematical model for immiscible gas-water displacement in the heterogeneous porous medium is implemented in COMSOL Multiphysics (R). The model successfully captures the dynamics of the injection pressure and fluid recovery with and without gravitational effects. The model is developed to predict the injection pressure corresponding to gas breakthrough. By injecting the gas phase at a pressure below this threshold value, the gas breakthrough can be significantly postponed. During this delayed time, more liquid recovery is expected. The model is verified with experiments using a constant injection rate scheme. We also study the optimal operation by considering recovery factor and process time as the variables, in horizontal gas injection case. Without the use of high capillary pressure skin and pressure control scheme, the ultimate recovery from porous medium is limited to that from the fracture only, (approximately 9% pore volume or PV). Using the skin and pressure control, the model predicts the recovery values to be increased by one order of magnitude, which is in agreement with the experimental data. The computational fluid dynamic (CFD) model can be used for process control and optimization to achieve optimal recovery conditions in remediation and enhanced oil recovery (EOR).