Multi-scale modeling framework

In the past several years, Michael Celia and colleagues have developed a series of analytical and semi-analytical models that allow simulation of CO2 injection and leakage with a focus on leakage along old wells. Geared toward quantitative estimation of leakage on large scales, the group’s methods are very efficient and allow for simulations that include hundreds to thousands of leaky wells across many layers of geological formations. While providing important practical insights into the leakage problem, the early models required a number of assumptions and simplifications.

The researchers have now developed a much broader range of simulation tools that eliminate many of the assumptions required in the earlier models. The current models rely on only one major assumption, which is vertical equilibrium for the pressures of both the CO2 and brine in each permeable formation. This assumption implies a time scale that is large relative to the time required for the fluids to segregate (by buoyancy) in the vertical direction. For many formations, where the formation thickness is small relative to the horizontal extent, and where injection operations will continue over many decades, this is a reasonable assumption.

With this single assumption, the models can be simplified significantly while still allowing for geometric and parameter complexity within a formation (Figure 14). The models currently include all important processes except complex geochemistry. Quantitative modeling of residual (capillary) trapping, dissolution including convective mixing, and transport and possible leakage of buoyant separate-phase CO2 are all included. The models also simulate brine movement and possible leakage and identify appropriate regions of elevated pressure associated with the “area of review.” In addition, the models include capillary-fringe representations of the transition zone between CO2 and brine; this fringe zone can have important impacts on both rates of CO2 plume movement and the shape of invading CO2 fronts.

These new models are solved with a combination of numerical and analytical methods. They are formulated in the context of a general multi-scale modeling framework, which is based on explicit estimation of important length and time scales. Whenever possible, scale separation is exploited in the models, and fine-scale resolutions are only included in local reconstructions or down-scalings. This allows calculations to be performed exclusively on large-scale grid blocks.

Figure 14. Modeled formation with different regions

In addition to model development, Celia and colleagues continue to work on detailed projects focused on well-leakage parameters in partnership with Andrew Duguid from Schlumberger (a former PhD student from Princeton). These include development of new software to more easily perform parameter estimations associated with data from Vertical Interference Tests (VIT) and an effort to optimize well-based observations to monitor for leakage of CO2 or brine.

Finally, the team has continued to develop a web interface that provides solutions for CO2 plume and pressure field evolution. They now have a module where leakage along a leaky well can be included in the simple web-based models. See

Active collaborations for the modeling activities include ongoing close collaborations with colleagues from the University of Bergen in Norway, especially Jan Nordbotten, as well as ongoing collaborations with Sarah Gasda (a Princeton PhD graduate) at the University of North Carolina. In addition, the team has a joint research project with the USEPA research center in Athens, Georgia to develop the hierachical modeling framework; a joint project with Mark Person at New Mexico Technical University to apply models to the Illinois Basin; and continuing collaborations with Stefan Bachu at Alberta Innovates (formerly Alberta Research Council). Finally, Michael Celia is a co-PI with Catherine Peters on a Department of Energy grant to Princeton entitled Basin-Scale Leakage Risks from Geologic Carbon Sequestration: Impact on CCS Energy Market Competitiveness.


  •  Michael Celia was 2010 ASCE Pioneers in Groundwater Lecturer (lecture focused on CO2 modeling work).
  • Celia continues to serve on the Scientific Advisory Board for the In Salah Project; on the Advisory Board for the CO2 Capture Project (Phase 3); and on the Organizing Committee for the IEA Wellbore Integrity Network (including the joint meeting in 2011 with the IEA Modeling Network to be held in Perth, Australia).
  • The Celia group organized and hosted a workshop at Princeton titled “Scales of Resolution, Model Complexity, and Solution Approaches for CO2 Storage Problems”
  • Post-doc Mark Dobossy (in conjunction with M. Celia, J. Nordbotten, and A. Janzen) has formed a start-up company, Geological Storage Consultants, to commercialize the injection simulation software. The company has received a small business grant from the Department of Energy for this work.