Impact of Carbon Dioxide Trapping on Geological Storage

If we are to avoid potentially dangerous climate change, we need to capture and store CO2 emitted by fossil-fuel burning power stations and other industrial plants [123]. Saline aquifers provide the largest potential for storage and the widest geographical spread [66]. Subsequent leakage of CO2 into the atmosphere, even over hundreds of years, would render any sequestration scheme inefficient. However, based on the experience of the oil and gas industry, there is a good understanding of trapping mechanisms that take place in geological formations. Carbon capture and storage (CCS), where carbon dioxide, CO2, is collected from industrial sources and injected underground is one way to mitigate atmospheric emissions of this major greenhouse gas (GHG). Possible sites to accommodate CO2 storage are saline aquifers and oil reservoirs. These two types of location are considered for two reasons: the enormous storage potential in aquifers and the additional hydrocarbon production that could be produced by oil reservoirs. It is important that the injection scheme is designed such that the CO2 is safely stored and will not escape to the surface. Residual trapping offers a potentially quick and effective alternative method by which a non-wetting phase is rendered immobile as recent modelling has suggested that up to 90% of CO2 can be effectively immobilised by residual trapping in a short (years to decades) timescale [133]. There are only a few experimental measurements of capillary trapping in unconsolidated media in the literature. This is because the experimental measurements of multi-phase flow are extremely difficult to perform and the results are frequently not reliable at low saturations [119]. Most of the studies concentrate on trapped gas and rather than the residual saturation of a liquid phase: CO2 stored underground will be super-critical and liquid-like. In this work, we focus on measuring reliably and precisely residual saturations for both two- and three-phase flow covering the entire saturation range, including very low residual saturations. We performed drainage-imbibition and buoyancy-driven experiments for two-phase flow (oil-water and gas-water systems) and three-phase gravity drainage experiments for an oil-gas-water system on unconsolidated sand (LV60). The measured porosity of the sand was 0.37 obtained from three replicates (each replicate is a completely new experiment). The mean absolute permeability was 3.1 x 10-11 m2. The initial water saturation (Swi), residual oil saturation (Sor) and residual gas saturation (Sgr) were measured by two methods, namely mass balance (MB) and volume balance (VB). Mean values were 0.27 for Swi, 0.13 for Sor, and 0.14 for Sgr. Accuracy was maintained to be within 0.1% for every measurement. The buoyancy-driven experiments results show that Sor and Sgr are 11% and 14% respectively and generally lower than consolidated media. The trapped saturations initially rise linearly with initial saturation to a maximum value, followed by a constant residual as the initial saturation increases further. This behaviour is not predicted by the most commonly-used empirical models, but is physically consistent with poorly consolidated media where most of the larger pores can easily be invaded at relatively low saturation and there is, overall, relatively little trapping. The best match to our experimental data was achieved with the trapping model proposed by Aissaoui [2]. The three-phase gravity drainage experiments results show that for high initial gas saturations more gas can be trapped in the presence of oil than in a two-phase (gaswater) system. This is unlike previous measurements on consolidated media, where the trapped gas saturation is either similar or lower to that reached in an equivalent twophase experiment. The maximum residual gas saturation is over 20%, compared to 14% for two-phase flow. For lower initial gas saturation, the amount of trapping follows the initial-residual trend seen in two-phase experiments, although some values lie below the two-phase correlation These results are discussed in relation to pore-scale displacement processes and compared to literature values – mainly on consolidated media – that find that both gas and oil residuals are lower in three-phase than twophase flow [32, 52, 70, 81, 95, 97, 101, 108, 143-145]. This work implies that CO2 injection in poorly consolidated media would lead to rather poor storage efficiencies, with at most 4-6% of the rock volume occupied by trapped CO2; this is at the lower end of the compilation of literature results shown in Fig. 5.2. Using the Land correlation to predict the behaviour would tend to over-estimate the degree of trapping except for high initial saturations. The presence of a third phase (such as in an oil field, for instance) may improve the trapping efficiency