Time-lapse imaging of flow instability and rock heterogeneity impacts on CO2 plume migration in meter long sandstone cores. (June 2022)

• Record Type: Journal Article
• Authors: Seyyedi, Mojtaba; Clennell, Michael Benedict; Jackson, Samuel J.
• Is Part Of: Advances in water resources. Volume 164(2022)
• Publication Date: 2022-06
• Subjects: CO2 storage -- Flow instability -- Rock heterogeneity -- Capillary heterogeneity -- Viscous -- Fingering; Hydrology -- Periodicals; Hydrodynamics -- Periodicals; Hydraulic engineering -- Periodicals; 551.48
• DOI: 10.1016/j.advwatres.2022.104216
• Languages: English
• ISSNs: 0309-1708

Abstract

Highlights: Gravitational forces in the absence of strong heterogeneity play the dominant role in gas plume migration. In the presence of strong heterogeneity, contrast in capillary entry pressure plays as important a role as gravitational forces. Abstract: Our understanding and modelling of flow instabilities (e.g. viscous, gravitational) in the presence of geological heterogeneity is limited by a lack of experimental observations at relevant scales in consolidated media. Typically, small sample sizes (<50 mm diameter) restrict the scale of heterogeneity, and the development of unstable fingers may be damped by the sample boundaries. Furthermore, there is a lack of 3D dynamic data, resolved at a level to observe finger growth and interaction. To this end, we conduct a series of drainage & imbibition fluid-flow experiments with nitrogen and brine in meter-long, 100 mm-diameter Bentheimer and Boise sandstone samples using an advanced, high-pressure setup with medical X-Ray CT (XCT) imaging. Using novel beam-hardening and noise removal methods, we can resolve fluid distributions with a spatial resolution of 1.2 mm 3 and temporal resolution of 90 s. The mobility ( M ) regime in each case is weakly unstable, M ≈ 0.5-5, with flow velocities equivalent to field scale injection rates. In the case of the weakly heterogeneous Bentheimer sample (porosity 0.26 ± 0.025), we find the plume migration is mainly controlled by buoyancy. On the other hand, the fluid distribution in the Boise sample is controlled by the strong unstructured heterogeneity (porosity 0.3 ± 0.1). The heterogeneity increases the gas trapping as less of the core is ultimately bypassed, and residual gas saturation increases behind local capillary barriers. We see little viscous instability in each experiment despite the unstable regime; structural heterogeneity and buoyancy largely control the flow under these field-relevant conditions. The meter scale 3D experimental dataset is provided open-access and represents an ideal benchmark to calibrate and improve multiphase flow models and upscaling methods.