Prediction of Plume behavior and leakage
In addition to complex numerical models, the group is also using fast analytical models that are capable of quickly estimating leakage rates for hundreds of wells. This year they built on previous work, adding pressure-dependent density and viscosity to their model of CO2 injection. Mike Celia and colleagues’ study indicates that the pressure-dependence can be neglected unless fluids are close to the critical point, so that models with constant properties provide reasonable estimates of subsurface injection behavior. They have also developed solutions for post-injection plume evolution, and for local upconing of brine around a well that is leaking CO2.
The team’s next step is to model a hypothetical injection at a site outside of Edmonton. The site is close to several major point sources of CO2 and has hundreds of potentially leaky wells. The storage group researchers are working with Stefan Bachu from Alberta Geological Survey to characterize the site, and will simulate a hypothetical injection in an area of 400 to perhaps a thousand square kilometers with both the analytical model and a coarse-grid simulation with upscaling.
The analytical model results will be used to develop criteria to assure that leakage will not exceed specified target rates. The idea is to use a series of Monte Carlo simulations to develop rating curves that relate overall leakage rates to the underlying statistics of the well properties. In this way, regulators can set maximum leakage rates, and operators would need to demonstrate compliance with given parametric values such as maximum mean permeability for all existing wells within the radius of influence of a proposed injection. This approach gives specific targets that need to be demonstrated by operators, and also leads to freedom to develop measurement methods to acquire data required to satisfy these criteria.