Continuum-scale convective mixing in geological CO2 sequestration in anisotropic and heterogeneous saline aquifers

Deep saline aquifers are important geological formations for CO2 sequestration. It has been known that dissolution of CO2 increases brine density, which results in downward density-driven convection and consequently greatly enhances CO2 sequestration. In this study, a continuum-scale lattice Boltzmann model is used to investigate convective mixing of CO2 in saline aquifers. It is found that increasing permeability in either the vertical or horizontal direction accelerates the development of convective mixing. In a heterogeneous aquifer, increasing heterogeneity hampers the onset of convective mixing, because the heterogeneous permeability field results in a large portion of low-velocity region which reduces the instability of the system. The critical time for the onset of instability depends mainly on the coefficient of variation (COV) of the permeability field, and is insensitive to the correlation length. This implies that within the scale of critical time, mass transport is dominated by diffusion, and thus depends mainly on fine-scale heterogeneity controlled by COV. We derived an empirical formula for estimating the critical time, which leads to good estimates for all combinations of COV and correlation length. Fingering, channeling, and dispersion are the three mechanisms for mass transport. In dispersion, dissolved mass is approximately proportional to the square root of time, while in fingering and channeling it is approximately proportional to time. Mass transport by channeling depends significantly on permeability structure, while by fingering it is controlled by gravitational instability. It is also found that larger volumes of CO2 can be stored in heterogeneous aquifers because of higher mass dissolution rates.

► LBM is used to study the continuum-scale convective mixing.
► Effects of heterogeneity on the onset of convective mixing is numerically studied.
► An empirical formula is given for estimating the critical time.