Simulations of subsurface multiphase flow including polymer flooding in oil reservoirs and infiltration in vadose zone
MetadataShow full item record
With the depletion of oil reserves and increase in oil price, the enhanced oil recovery methods such as polymer flooding to increase oil production from water flooded fields are becoming more attractive. Effective design of these processes is challenging because the polymer chemistry has a strong effect on reaction and fluid rheology, which in turn has a strong effect on fluid transport. We have implemented a well-established polymer model within the Implicit Parallel Accurate Reservoir Simulator (IPARS), which enables parallel simulation of non-Newtonian fluid flow through porous media. The following properties of polymer solution are modeled in this work: 1) polymer adsorption; 2) polymer viscosity as a function of salinity, hardness, polymer concentration, and shear rate; 3) permeability reduction; 4) inaccessible pore volume. IPARS enables field-scale polymer flooding simulation with its parallel computation capability. In this thesis, several numerical examples are presented. The result of polymer module is verified by UTCHEM, a three-dimensional chemical flood simulator developed at the University of Texas at Austin. The parallel capability is also tested. The influence of different shear rate calculations is investigated in homogeneous and heterogeneous reservoirs. We observed that the wellbore velocity calculation instead of Darcy velocity reduces the grid effect for coarse mesh. We noted that the injection bottom hole pressure is very sensitive to the shear rate calculation. However, cumulative oil recovery and overall oil saturation appear to not be sensitive to grid and shear rate calculation for same reservoir. There are two models to model the ground water infiltration in vadose zone. One is Richard’s Equation (RE) model. And the other is two-phase flow model. In this work, we compare the two-phase model with an RE model to ascertain, under common scenarios such as infiltration or injection of water into initially dry soils, the similarities and differences in solutions behaviors, the ability of each model to simulate such infiltration processes under realistic scenarios, and to investigate the numerical efficiencies and difficulties which arise in these models. Six different data sets were assembled as benchmark infiltration problems in the unsaturated zone. The comparison shows that two-phase model holds for general porous media and is not limited by several assumptions that must be made for the RE formulation, while RE is applicable only for shallow regions (vadose) that are only several meters in depth and a fully saturated bottom boundary condition must be assumed.