HR: 1340h
AN: GC13A-1215    [Abstracts]
TI: Quantitative Estimation of Leakage Potential in Mature Sedimentary Basins
AU: * Celia, M A
EM: celia@princeton.edu
AF: Princeton University, Department of Civil and Environmental Engineering, Princeton, NJ 08544 United States
AU: Bachu, S
EM: Stefan.Bachu@gov.ab.ca
AF: Alberta Geological Survey, Alberta Energy and Utilities Board, Edmonton, AB T6B 2X3 Canada
AU: Nordbotten, J M
EM: janmn@mi.uib.no
AF: Princeton University, Department of Civil and Environmental Engineering, Princeton, NJ 08544 United States
AU: Nordbotten, J M
EM: janmn@mi.uib.no
AF: University of Bergen, Department of Mathematics, Bergen, N-5020 Norway
AU: Kavetski, D
EM: kavetski@princeton.edu
AF: Princeton University, Department of Civil and Environmental Engineering, Princeton, NJ 08544 United States
AU: Gasda, S E
EM: sgasda@princeton.edu
AF: Princeton University, Department of Civil and Environmental Engineering, Princeton, NJ 08544 United States
AB: Many deep saline aquifers suitable for CO2 injection are located in mature, onshore sedimentary basins of North America. These basins have been subjected to more than a century of oil and gas exploration and production, as well as the more recent practice of deep waste disposal. This legacy of drilling has resulted in large numbers of wells, a significant fraction of which are abandoned. For example, the number of wells is approaching 400,000 in the Alberta Basin in Canada, while in Texas the number exceeds 1,000,000. Because these wells serve as conduits along which injected CO2 could leak, any risk assessment focused on large-scale CO2 injection and storage should include an analysis of leakage potential along these wells. We have analyzed spatial statistics of wells in a particular formation in the Alberta Basin, and found densities as high as 5 wells per km2. This implies that in developed parts of the basin, where infrastructure exists for injection operations, a typical CO2 injection plume may contact hundreds of existing wells. Because data on existing wells, especially abandoned wells, are very scarce, there is high uncertainty associated with quantitative measures of well properties and parameters. Therefore a systematic analysis of leakage potential would necessarily include a Monte Carlo type of approach involving many realizations of leakage simulations. In order to be able to perform Monte Carlo simulations under conditions of high uncertainty, we have developed a semi-analytical model for CO2 injection and associated leakage in fields that have arbitrary numbers and locations of wells, over domains that include vertical multi-layered successions of aquifers and aquicludes. The efficiency of this model allows us to solve hundreds to thousands of Monte Carlo realizations for any assigned probability distributions for well properties, thereby providing quantitative relationships between well parameters, injection conditions, and leakage rates. In this presentation, we will give an overview of our semi-analytical model and then present a specific example field application. The field application is based on data at a location in Alberta, close to several large point emissions of CO2. We have run the model over a domain of 30km x 30km, which includes more than 500 existing wells. By setting as constant all parameters in the model except the effective permeability in the existing wells, we can determine the probability distribution of total leakage rates as a function of the parameters (means and variances) of the permeability distribution. By sampling different probability distributions, we can identify threshold statistics associated with leakage targets. We will use these results to demonstrate how the model works, how it fits into a general statistical framework, and how it fits into an overall risk analysis of CO2 injection and storage in brine aquifers located in mature sedimentary basins.
DE: 1828 Groundwater hydraulics
DE: 1849 Numerical approximations and analysis
DE: 1859 Rocks: physical properties
SC: Global Climate Change [GC]
MN: Fall Meeting 2005