Development and calibration of a reactive transport model for carbonate reservoir porosity and permeability changes based on CO2 core-flood experiments

Beneficial pore space and permeability enhancements are likely to occur as CO2-charged fluids partially dissolve carbonate minerals in carbonate reservoir formations used for geologic CO2 storage. The ability to forecast the extent and impact of changes in porosity and permeability will aid geologic CO2 storage operations and lower uncertainty in estimates of long-term storage capacity. Our work is directed toward developing calibrated reactive transport models that more accurately capture the chemical impacts of CO2-fluid-rock interactions and their effects on porosity and permeability by matching pressure, fluid chemistry, and dissolution features that developed as a result of reaction with CO2-acidified brines at representative reservoir conditions. We present new results from experiments conducted on seven core samples from the Arbuckle Dolostone (near Wellington, Kansas, USA, recovered as part of the South-Central Kansas CO2 Demonstration). Cores were obtained from both target reservoir and lower permeability baffle zones, and together these samples span over 3-4 orders of magnitude of permeability according to downhole measurements. Core samples were nondestructively imaged by X-ray computed tomography and the resulting characterization data were mapped onto a continuum domain to further develop a reactive transport model for a range of mineral and physical heterogeneity. We combine these new results with those from previous experimental studies (Smith et al., 2013; Hao et al., 2013) to more fully constrain the governing equations used in reactive transport models to better estimate the transition of enhanced oil recovery operations to long-term geology CO2 storage. Calcite and dolomite kinetic rate constants (mol m(-2) s(-1)) derived by fitting the results from core-flood experiments range from k(calcite,25c) = 10(-6.8) to 10(-4.6), and k(dolomite),(2SC) = 10(-7.5) to 10(-5.3). The power law-based porosity-permeability relationship is sensitive to the overall pore space heterogeneity of each core. Stable dissolution fronts observed in the more homogeneous dolostones could be accurately simulated using an exponential value of n = 3. Unstable dissolution fronts consisting of preferential flowpaths could be simulated using an exponential value of n = 3 for heterogeneous dolostones, and larger values (n = 6-8) for heterogeneous limestones. (C) 2017 The Authors. Published by Elsevier Ltd.