Laboratory Flow Experiments for Visualizing Carbon Dioxide-Induced, Density-Driven Brine Convection

Injection of carbon dioxide (CO 2 ) into saline aquifers confined by low- permeability cap rock will result in a layer of CO 2 overlying the brine. Dissolution of CO 2 into the brine increases the brine density, resulting in an unstable situation in which more-dense brine overlies less-dense brine....

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Transport in porous media 2010-03, Vol.82 (1), p.123-139
Hauptverfasser: Kneafsey, Timothy J., Pruess, Karsten
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Injection of carbon dioxide (CO 2 ) into saline aquifers confined by low- permeability cap rock will result in a layer of CO 2 overlying the brine. Dissolution of CO 2 into the brine increases the brine density, resulting in an unstable situation in which more-dense brine overlies less-dense brine. This gravitational instability could give rise to density-driven convection of the fluid, which is a favorable process of practical interest for CO 2 storage security because it accelerates the transfer of buoyant CO 2 into the aqueous phase, where it is no longer subject to an upward buoyant drive. Laboratory flow visualization tests in transparent Hele-Shaw cells have been performed to elucidate the processes and rates of this CO 2 solute-driven convection (CSC). Upon introduction of CO 2 into the system, a layer of CO 2 -laden brine forms at the CO 2 -water interface. Subsequently, small convective fingers form, which coalesce, broaden, and penetrate into the test cell. Images and time-series data of finger lengths and wavelengths are presented. Observed CO 2 uptake of the convection system indicates that the CO 2 dissolution rate is approximately constant for each test and is far greater than expected for a diffusion-only scenario. Numerical simulations of our system show good agreement with the experiments for onset time of convection and advancement of convective fingers. There are differences as well, the most prominent being the absence of cell-scale convection in the numerical simulations. This cell-scale convection observed in the experiments may be an artifact of a small temperature gradient induced by the cell illumination.
ISSN:0169-3913
1573-1634
DOI:10.1007/s11242-009-9482-2