SPE Journal
Volume 13, Number 3, September 2008, pp. 305-313

Summary

The problem of CO₂ sequestration in geologic formations is analyzed from a fundamental perspective. In order to clearly understand the first order behavior of the system, the mechanisms of trapping, dissolution and chemical reactions are not accounted for. The analysis is concerned with the post-injection period when the CO₂ plume rises due to buoyancy. Characteristics of the plume for a 1D problem show that a pair of shocks moving in opposite directions is produced at the top end. The downward moving shock interacts with the bottom end of the plume resulting in a decrease in the maximum value of the CO₂ saturation. High accuracy numerical simulations are employed to understand the 2D mechanisms of plume evolution in terms of the viscosity ratio and the capillary number. 2D results show that the plume rises to significantly lower depth,in shorter times, as compared to the 1D problem. This behavior is governed by the 2D velocity field around the plume that additionally leads to spanwise wave interactions and results in a faster decrease of the maximum CO₂ saturation. The initial dimensions of the plume have a strong influence on the time scales of the wave interactions. The maximum upward velocity that is generated due to buoyancy is closely related to the maximum saturation and decays rapidly to very small values with a decrease in saturation. In the case where the viscosity of CO₂ is a tenth of the viscosity of the surrounding fluid, the plume rises up about 500 m in 700 yrs. Our results provide an upper bound on the maximum rise distance and the sequestration time for the problem involving trapping and dissolution. Comparison with experimental results show that the buoyancy velocity obtained from our results is of the same order as observed in the experiments.

Introduction

The behavior of two-phase flow in porous media under conditions of unstable density stratification is an important and challenging problem applicable to many practical settings of interest. Particularly,the dynamics of two-phase immiscible flows that are gravitationally unstable play a central role in the area of carbon dioxide storage in depleted reservoirs and saline aquifers. The important issue in this regard is the understanding and prediction of the fate of CO₂ over a time period of geological scale (Bachu et al. 1994). The success of CO₂ sequestration operations in subsurface geological formations is critically linked to the ability of the storage site to sequester the gas indefinitely. The main mechanisms of sequestration are microscopic residual trapping, dissolution of gas into brine, and chemical fixing of carbon into the rock (Gunter et al. 1997). Various time scales as well as the nature of the storage site determine the relative importance of these mechanisms.

The sequestration process can be broadly classified into three phases (Ennis-King and Paterson 2005). Namely, the injection phase where super-critical CO₂ is injected into the site. This is followed by the post-injection period where the gas rises as a buoyant plume. Residual trapping and dissolution will be of primary importance in this stage. The final stage is thought to be governed by dissolution driven gravitationally unstable flows (Riaz et al. 2006) as well as chemical reactions of CO₂ with the porous rock. During the initial stage, the density and viscosity of the injected gas are less than the resident brine; therefore, the flow can potentially become unstable hydrodynamically (Riaz and Tchelepi 2007) due to unfavorable contrasts of density and viscosity. The extent of the initial injection period is determined by the amount of carbon dioxide that can be stored in a given reservoir; however, this period is expected to be much shorter than the subsequent post injection period. The modeling of the injection process, which is based on the relative permeability formulation of the Darcy equations, is thought to be well developed for drainage type displacements that occur during this phase (Kumar et al. 2005). Most of the main trapping mechanisms are either unavailable or are relatively less important during the initial period. For example, residual trapping does not take place during drainage while chemical trapping occurs over much larger time scales. Viscous instability at the macroscopic scale is also a possibility (Riaz and Tchelepi 2006).

Our understanding of the evolution of the CO₂ plume during the post injection period is incomplete. It is during this stage that the processes of residual trapping and dissolution are expected to play a primary role. While in general one can expect the plume to rise due to buoyancy, the particular mechanisms of transport are in the initial stages of investigation (Wood et al. 2004; Stohr and Khalili 2006; Tokunaga et al. 2000). For example, what is the most appropriate model for the flow; is the relative permeability model appropriate, or given the extremely small capillary numbers, should the invasion percolation model be used (Yortsos et al. 2001)? Regardless of which model is used, the main questions that need to be addressed are: how far can the plume rise; what is the velocity of rise, and how far does the plume spread during its ascent? The last issue is important from the point of view of dissolution which occurs immediately when the gas comes into contact with unsaturated brine. However, because the saturation threshold of brine is small, a continuous supply of fresh brine around the buoyant plume can increase dissolution significantly. In this investigation we attempt to understand the dynamics of the CO₂ plume during the period immediately following the injection phase. We use the Darcy relative permeability model to analyze the dynamics governing the natural convection of the buoyant plume and provide some preliminary estimates of how far and how fast will the plume rise. In order to focus on the primary characteristics of transport governed by the Darcy model, we carry out the analysis for homogeneous rocks, without residual trapping and dissolution. Hence, the first order behavior of the system will be considered as a primary guide to subsequently develop a better understanding of more complex processes. A sketch of the CO₂ plume is shown in Fig. 1. The non-wetting gas phase is immersed in brine which is the wetting phase. A buoyancy force per unit volume F_B results in upward motion with velocity proportional to U_B ,inducing a downward flow of brine around the plume.

© 2008. Society of Petroleum Engineers

View full textPDF ( 448 KB )

History

• Original manuscript received: 28 June 2006
• Meeting paper published: 24 September 2006
• Revised manuscript received: 8 November 2007
• Manuscript approved: 12 November 2007
• Version of record: 20 September 2008