Multicomponent Inhomogeneous Fluid Transport in Low Permeability Oil Reservoirs

Abstract

Abstract Various transport mechanisms and phenomena unique to nanopores influence oil production from low permeability reservoirs, such as shales. One such phenomenon is the inhomogeneity of fluid properties across a pore width due to the confinement and pore wall effects. We propose a multicomponent fluid transport model for oil production from shale reservoirs by considering inhomogeneous fluid thermodynamics and transport properties based on pore-scale density distribution. We adopt the multicomponent simplified local density (MSLD) method incorporating fluid-fluid and fluid-solid interaction through the Peng-Robinson equation of state (PR-EOS) and 10-4 Lennard-Jones fluid-wall potentials to calculate density profiles in slit nanopores. Viscosity and diffusivity profiles are calculated based on the density profile. We solve a multicomponent momentum balance equation combined with the Maxwell-Stefan equation to obtain velocity profiles. We then use the area-averaged transmissibility in the multicomponent transport model based on the Maxwell-Stefan theory to simulate co- and counter-diffusion processes mimicking oil production and solvent (gas) injection processes. In addition to using the MSLD method, we employ PR-EOS and modified PR-EOS (with critical parameters shifts), representing homogenous fluid systems without and with confinement effects, to calculate thermodynamics and transport properties at pore- and continuum-scale. Porescale investigation results for a ternary hydrocarbon mixture (methane, propane, n-octane) within shale nanopores reveal that, in the case of hydrocarbon distribution in organic slit nanopores, the heaviest component exhibits a notable preference for the near-wall region due to pronounced fluid-solid interaction, while the composition in the pore-center region resembles that of the bulk fluid. Transport of the heavy component (n-octane) is enhanced at the near-wall region with a width approximately 1.5 times the fluid molecular collision diameter. Based on the deviation of the averaged mass flux ratio from unity, the pore size can be categorized into three fluid systems: inhomogeneity dominant (da < 3 nm), transition (3 nm < da < 30 nm), and homogeneity dominant (da > 30 nm) system. The fluid-wall interaction can be neglected in pores larger than 30 nm. However, fluid-solid interaction becomes increasingly significant as pores become smaller. Continuum-scale co-diffusion and counter-diffusion simulations show that, in the inhomogeneity dominant fluid system, neglecting the influence of inhomogeneous fluid results in a more than 30% overestimation of cumulative production/injection. Conversely, in the homogeneity dominant fluid systems, the impact of inhomogeneous fluid can be disregarded as the difference in cumulative production/injection is less than 1%. Furthermore, the results reveal that the commonly used modified PR-EOS incorporating critical parameter shift increases the errors associated with cumulative production and injection, resulting in even larger discrepancies between predicted and actual production/injection values. Therefore, when the fluid-wall interaction parameters are unavailable or the numerical simulations require excessive computational resources, it is advisable to utilize the PR-EOS instead of a modified PR-EOS to calculate transport coefficients and simulate fluid transport in low permeability reservoirs.