Reservoir Modeling of CO2 Plume Behavior Calibrated Against Monitoring Data From Sleipner, Norway

Abstract

Sleipner is a commercial CO2 storage site in the North Sea with good constraints from monitoring data, but also with some significant uncertainties regarding temperature, pressure and gas/brine behavior. At Sleipner, we have used high-quality repeated seismic and gravimetric surveys for monitoring and calibrating the reservoir uncertainties. To model the CO2 behavior we have used two main approaches: a) traditional reservoir simulations, using black oil and compositional fluid descriptions; and b) invasion percolation simulations, using threshold pressure and fluid density descriptions that assume the dominance of capillary and gravity forces. The key findings from the study are: The invasion percolation simulation gave the best initial match to observed data, leading us to reassess the input assumptions for the black oil and compositional simulations. By taking into account gravity segregation and modifying the reservoir simulation input data, we were able to get a much better match for the black oil and compositional simulations. There is still scope for further optimization and history matching, however, this study has reduced the range of domain variables leading to an improved understanding of the flow processes involved in geological storage of CO2 in saline formations. The study has led us to conclude that we can make realistic and predictive CO2 storage models provided that the site-specific conditions are honored, including reservoir and fluid property characterization. The necessary tight constraints on input parameters are achieved by calibration against monitoring data. Our study illustrates both a rather novel approach to modeling CO2 storage and the need for improved input to conventional simulators. Application of our approach to other CO2 storage sites will help in achieving more realistic understanding of CO2 storage, thereby contributing to the maturation of CO2 storage technology worldwide.

Numerical Simulation and Optimization of Sleipner Carbon Sequestration Project

“…However, as higher CO 2 column accumulates, the pressure gradient gradually increases and leads to increasing breakthrough mass flux. As depicted in Figure 6, excellent agreement for the ten-year CO 2 flux analysis for the topmost sandstone layer is obtained between our simulation and the seismic amplitudes analysis [23] suggesting the overall accuracy of our modeling despite some discrepancy at detailed level. …”

“…The topmost sandstone layer of Utsira formation near the injection site, Layer #9, is of most interest since it has the highest concentration of gaseous CO 2 and has direct contact with the overlying caprock formation. Seismic survey has shown striking growth of CO 2 accumulation in Layer #9 between 1999 and 2006 as shown in Figure 11 [23]. The black dot in Figure 11 marks the location of the injection well, which is roughly 200 m under Layer #9.…”

Spill-point analysis and structural trapping capacity in saline aquifers using MRST-co2lab

“…In particular, we have outlined the algorithms used, demonstrated their sensitivity to geometry variations, and demonstrated how to identify injection points with the largest potential for structural trapping. In general, we believe that much of the information needed to optimize a CO 2 injection scenario can be obtained using simplified tools that honor the main dynamics in the system, gravity flow, as also noticed by [18], who used percolation type of calculations. In [4,5,6], we demonstrate how the tools described herein can be combined with vertical-equilibrium models and rigorous mathematical optimization to develop scenarios for the injection of hundreds of megatonnes of CO 2 into saline aquifers in the North Sea.…”

Reexamining CO2 Storage Capacity and Utilization of the Utsira Formation

“…In the present work, we use the atlas model of Utsira together with parameters from (Singh et al, 2010;Holloway et al, 2004;Chadwick et al, 2008), as listed in Table 11 to estimate volumetric trapping capacities. These parameters suggest a temperature at the depth of the current injection point of 40.2 • C, in the middle of the suggested range of 36 • C to 46 • C from (Chadwick et al, 2008).…”