270891 Reducing Uncertainties in Model Predictions Via History Matching of CO2 Plume Migration At the Sleipner Project, Norwegian North Sea

Tuesday, October 30, 2012
Hall B (Convention Center )
Peng Lu1, Chen Zhu1 and Per Aaggard2, (1)Department of Geological Sciences, Indiana University, Bloomington, Bloomington, IN, (2)Department of Geosciences, University of Oslo, Oslo, Norway

The Sleipner Project in Norway is the world’s first industrial-scale geological carbon sequestration project. Starting in 1996, CO2 separated from natural gas has been injected into the Utsira Sand at the rate of approximately 1 million metric ton of CO2 per year. To date, more than ~15 Mt of CO2 has been injected. Seismic surveys of the site were conducted prior to injection in 1994, and then repeated in 1999, 2001, 2004, 2006, and 2008; 2010 survey data have not yet been released by StatOil. These surveys have produced high fidelity 4D seismic data that delineated the CO2 plume migration history. Therefore, the Sleipner Project provides a somewhat unique opportunity to simulate the dynamics of CO2 plume migration in a real geological system, which in turn gives insights into the uncertainties in CO2plume migration prediction. The results will complement the numerous “concept models” proliferated in the literature, which use ideal geometry and homogenous geological properties.

Through a project sponsored by the U.S. Department of Energy/National Energy Technology Laboratory's Carbon Storage Program, we have developed a multi-phase reactive flow reservoir model of Layer 9 of the Utsira Sand for the Sleipner project using the Computer Modeling Group’s reservoir simulator GEM®. We calibrated the model against the time-lapsed seismic monitoring data. Our simulation results match with the extents of CO2 plume migration history from 1999 to 2008. The successful match with historic plume development was aided by introducing permeability anistropy and a second feeder to Layer 9. Predicted gas saturation, thickness of the CO2 accumulation and CO2 solubility in brine are also comparable with interpretations of the seismic data in the literature. The model calculated that ~9% of total injected CO2 is dissolved in brine, which is comparable to estimates (5-10%) based on seismic data interpretation. Our reservoir model was based on StatOil’s geological model of the Utsira formation and grid mesh. Our simulation results illustrate that the actual behaviors of the injected CO2 plume conform to the modeled behaviors.            

The Sleipner project is on-going. The good match of plume history provides a calibrated model for making predictions of CO2plume migration into the future. By comparing the model prediction and monitoring data that are not used in the calibration, we will have well-grounded assessment of modeling uncertainties.


Although the work was partly sponsored by an agency of the United States Government, the views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

Extended Abstract: File Uploaded