Study of reactive flow, ground deformation and real-time tomography with applications on CO₂ sequestration

Abstract

Geological sequestration of CO₂ is an option to either mitigate or defer global warming and avoid dangerous climate change. Carbon dioxide is also injected into reservoirs to increase resource extraction. Currently, the rate of CO₂ injection in pilot sequestration plants or the enhanced oil recovery commercial projects is a few million tons CO₂ per year at best, while there are tens of billion tons of annual carbon dioxide emissions. These sequestration projects will have a tangible impact only if they can be scaled up, which will require a solid understanding of the underlying physics and a portfolio of monitoring tools to curb operation risks and the potential of leakage. In this thesis I describe work that bears on three aspects of these complicated issues: Experiments and a new computational model on the reactive flow of carbonic acid in limestone, an assessment of the surface uplift owing to sequestration of CO₂ in a carbonate saline aquifer, and a validation of a new, real-time, tomography method that monitors the velocity change resulting from hydraulic fracturing. In Chapter 2, a network approach that models the formation of wormholes from reactive flow of high concentration carbonic acid in limestone is devised. In this model, the pore space is partitioned into two parts, the bigger subset of the pore network as a leading sub-network while the reminder of porosity comprises a set of identical secondary sub-networks: the reactive flow problem is framed as the competition of reactive fluid among these sub-networks. This approach saves computational resources by approximating the large fraction of slowly changing part of the pore space as a number of identical coarse networks. Using material constants appropriate to the conditions of the experiments, the model successfully grows wormholes that are as large as a fraction of a millimeter big in diameter from micron scale heterogeneity, and matches results obtained in laboratory experiments. In Chapter 3, the surface deformation from the injection of high pressure CO₂ in Nisku carbonate formation in Alberta, Canada is simulated with a coupled model of two-phase flow and geomechanics. The fact that the injected CO₂ is dry and the analyses of the flow field predict upper bounds of 2 percent stiffness decrease and 2 fold permeability increase when the injection rate in the horizontal pipe is 0.2 Mt/km/yr. Simulation with a constitutive law that incorporates these estimates shows the difference between the surface uplift from the injection in the carbonate reservoir and the poroelastic response are negligible given the resolution of geodetic observations. A simple poroelastic model is adequate to infer the migration of subsurface CO₂ for carbonate reservoirs as tight as the Nisku formation for the first five years of injection. In Chapter 4, a real time tomography method that monitors the change of velocity structure from hydraulic fracturing is presented. A data set is synthesized from forward simulation of the simultaneous occurrence of micro-seismic events and evolution of velocity structure. The new method adapted from Algebraic Reconstruction Technique successfully estimated the changes in the contour and extent of the low-velocity fracture zone with time.