Content area
Abstract
A finite element model of two-phase flow of air and water movement through porous media was developed. The formulation for radial flow used axisymmetric linear triangular elements. Due to the radial nature of the problem, a two dimensional formulation was used to represent three dimensional space. Governing equations were based on Darcy's equation and continuity. Air was treated as a compressible fluid by using the Ideal Gas Law.
A gravity-driven saturated flow problem was modeled and the predicted flow rate exactly matched the analytical solution. Comparisons of analytical and experimental results of one-phase radial and vertical flow were made in which capillary pressure distributions were almost exactly matched by the two-phase model (TPM). The effect of air compression on infiltration was simulated. It was concluded that the TPM modeled air compression and its inhibiting effect on infiltration even though air counterflow through the surface boundary was not permitted. The difficulty in describing the boundary conditions for air at a boundary where infiltration occurred was examined.
The effect of erroneous input data for the soil moisture characteristic curve and the relative permeability curve was examined. It was determined that capillary pressure varied significantly with small differences in the soil moisture characteristic curve, and therefore, accurate data are a necessary input. It was also shown that characteristic curve data that was outside the range of capillary pressures to which the TPM was applied had a detrimental effect on capillary pressure prediction.
An attempt was made to simulate the long term effects of air injection into the unsaturated zone above a water table within an aquifer. The modeling of the air - water table interface was unsuccessful. The cause of the problem was determined to be the modeling of the air phase, which becomes discontinuous at the air - water table interface, by continuum mathematics. Solutions regarding alternative formulations to circumvent the problem of a discontinuous air phase in a saturated zone were presented. It was concluded that the TPM would model two-phase situations but cases where a standing water table existed could not be simulated.
