Use of pore-scale modeling to understand flow and transport in porous media

Multiphase flow and transport processes in porous media are involved in a wide variety of engineering applications, such as groundwater remediation and oil recovery. The inherent heterogeneity of subsurface porous media, as well as the complexity involved in the multiphase physics, result in a significant challenge to the fundamental understanding of multiphase flow and transport. In this work, we study flow and transport problems using pore-scale modeling approaches, which allow us to investigate microscale processes and their effect on macroscale behavior.; Two pore-scale approaches, lattice-Boltzmann (LB) and pore-network modeling, are used to study flow in simulated sphere packings that vary in porosity and sphere-size distribution. For both modeling approaches, the size of the representative elementary volume is determined with respect to the porous media permeability. Permeabilities obtained by LB simulations agree well with empirical permeability relations in sphere packings. Pore-network simulations underestimate the permeability due to the simplified representation of the porous media. Based on LB simulations in packings with log-normal sphere-size distribution, we suggest a constitutive relation for permeability as a function of porosity, as well as the mean and standard deviation of the sphere diameter.; We have also developed a multiphase LB model to investigate the concurrent flow dynamics in porous media. A model two-fluid-phase sphere-pack system is designed to mimic laboratory experiments conducted to evaluate the hysteretic capillary pressure saturation relation for a system consisting of water, tetrachloroethylene, and a glass bead porous medium. Good agreement is achieved between the measured hysteretic capillary pressure saturation relations and the LB simulations when comparing entry pressure, displacement slopes, irreducible saturation, and residual entrapment. We also suggest a lower bound on the size of a representative elementary volume with respect to the capillary pressure saturation relations.; The LB method is a powerful approach, but it is also computationally demanding, requiring the use of massively parallel computing approaches. Conventional LB implementations for simulating flow in porous media store the full lattice, making parallelization straightforward but wasteful in floating point operations and storage. We develop a new two-stage parallel implementation consisting of a sparse domain decomposition stage and a simulation stage that avoids the need to store and operate on lattice points located within the solid phase. Results indicate that the new LB implementation combined with orthogonal recursive bisection (ORB) decomposition is a promising approach, showing near linear scaling and requiring substantially less storage and computational time compared to conventional LB approaches.