Streamline Based Simulation of CO2 Injection in ...

Object Details

View

XVI International Conference on Computational Methods in Water Resources (CMWR-XVI) Ingeniørhuset

Streamline Based Simulation of CO2 Injection in Saline Aquifers
Paper
Author:Kristian Jessen <krisj@stanford.edu> (Stanford University)
Hamdi Tchelepi <tchelepi@stanford.edu> (Stanford University)
Franklin M. Orr, Jr. <fmorr@stanford.edu> (Stanford University)
Presenter:Kristian Jessen <krisj@stanford.edu> (Stanford University)
Date: 2006-06-18     Track: Special Sessions     Session: Geologic Sequestration of Carbon Dioxide
DOI:10.4122/1.1000000358
DOI:10.4122/1.1000000359

Design and implementation of CO2 sequestration projects in saline aquifers require, in part, a solid understanding of the key physics that determine distribution of the injected CO2 within the target aquifer. This understanding then forms the basis for model formulation and development of simulation techniques appropriate for resolving the essential physics. In addition, a significant level of uncertainty exists as to the spatial distribution of rock properties. This, in turn, calls for development of simulation tools that are accurate and sufficiently efficient to be included in uncertainty estimation frameworks. In this paper we propose compositional streamline simulation as a viable method of predicting the movement of CO2 in aquifers during the injection period. We demonstrate how to handle the compositional effects (solubility of CO2 in brine) in a very efficient manner based on look-up tables and explicit calculation of phase distributions, phase compositions and transport properties. The compositional streamline approach is compared to traditional finite difference/volume simulation techniques based on explicit, adaptive implicit and fully implicit formulations. The comparison includes compositional and equivalent “black oil” formulations to assess the efficiency and accuracy of existing technology relative to compositional streamline simulation. We demonstrate the proposed formulation to be significantly less CPU intensive than existing methods for most flow settings. For larger scale problems with more than 500K active grid cells, the run times for the proposed formulation are more than an order of magnitude lower than those for conventional numerical approaches. We conclude by discussing the limitations of the streamline method for post- injection time scales, where gravity-driven flows, diffusion of dissolved CO2, residual trapping of CO2 due to flow reversal and CO2 saturation reduction as a result of additional dissolution, and geochemistry become important mechanisms determining the long term fate of injected CO2. Computational methods other than the streamline approach are more appropriate for the post-injection period.