H52F-05
System response as a function of relative permeability in geologic CO2 sequestration

Friday, 18 December 2015: 11:20
3018 (Moscone West)
Ryan Pollyea, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States
Abstract:
Within the portfolio of risk assessment strategies for carbon capture and sequestration projects, numerical modeling and simulation is frequently utilized for predicting CO2 storage capacity, leakage potential, and geomechanical reservoir integrity. In numerical CO2 injection models, one commonly used approach for simulating the effects of relative permeability (krel) is to apply the van Genuchten (1980) model to the wetting phase and the Corey (1954) model to the non-wetting phase. In this formulation, wetting phase permeability decay is controlled by a phase interference parameter (m), and non-wetting phase permeability decay is controlled by irreducible non-wetting phase saturation (Sgr). Although krel is a well-known phenomenon, there exists much uncertainty in parameterizing krel models and little is known about the influence of parameter space on reservoir performance. This work presents results from a numerical modeling experiment designed to isolate the effects of variability in krel parameters, m and Sgr. A series of CO2 injection simulations is performed for 399 unique combinations of m and Sgr, which vary systematically over a range of 0.1 – 0.99 and 0.01 – 0.50, respectively. CO2 is injected at modest 2.78 kg/s for 10 years into a radially symmetric grid using a beta version of TOUGH3/ECO2N, and all reservoir properties remain constant across the ensemble of 399 simulations. Results from this work show the injection pressure response ranging from ~10 MPa to >60 MPa, where the high end of this range is focused on a narrow portion of the parameter space corresponding to m < 0.3 for all Sgr. Additionally, the maximum CO2 plume radius (defined as CO2 saturation > 0.01) ranges from 550 m to 1250 m, where the high end of this range corresponds with low m and low Sgr. Although the model geometry is quite simple, these results demonstrate enormous variability in the both injection pressure and CO2 plume dimension solely as a function of krel.