Validation of a carbonate-brine-CO2 reactive transport model against new experimental core-flooding results

Subsurface carbonate formations are of special interest for geologic carbon sequestration (GCS) because their characteristically irregular pore geometries and fast mineral dissolution rates in the presence of CO₂ may result in more rapid changes to void space structures and local permeability. Given this potential, CO₂-EOR and GCS operations in carbonate reservoirs could benefit from better reactive transport estimates for potential long-term changes to CO₂ reservoir porosity and permeability. In previous work we have developed and calibrated one such reactive transport model by matching pressure, single-phase solution chemistry, and dissolution features from fifteen core-flood experiments from two different carbonate reservoirs (Weyburn-Midale carbonate unit, Saskatchewan, Canada; Arbuckle dolomite, Kansas, USA). Here we validate this model with a coupled experimental and modeling study using samples from a third distinct reservoir, the Duperow formation (Montana, USA). Forward simulations are compared with experimental results from five new experiments. Simulations sample the parameter ranges for calcite and dolomite rate constants (log k_cal,25C = -6.5, -5.5, -4.5; and log k_dol,25C = -8, -7, -6)), and for the porosity-permeability relationship exponent n (1.5, 3, and 8). Model domains honor the initial bulk permeability of each sample, as well as the mineral and void space distributions determined from non-destructive characterization prior to CO₂-brine reaction. The simulated forecast envelopes generated by the reactive transport model capture the pre-breakthrough porosity generation trends and declines in pressure declines observed from these new experimental datasets. This validation study demonstrates that model input parameters can be used to estimate changes to new void space generation and bulk permeability, at the core scale within experimental uncertainty. The reactive transport model predicts changes to void space to within a factor of 40, while permeability forecasts can range by orders of magnitude depending on subsurface heterogeneity and the value of the porosity-permeability exponent used in the simulations. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.