Application of arbitrary polynomial chaos (aPC) expansion for global sensitivity analysis of mineral dissolution and precipitation modeling under geologic carbon storage conditions

Abstract

Numerical modeling of geochemistry associated with geologic CO2 storage involves many conceptual and quantitative uncertainties. In this study, a time efficient arbitrary polynomial chaos (aPC) expansion approach was proposed to do global sensitivity analysis of mineral dissolution and precipitation modeling in geologic carbon storage scenarios. To demonstrate the workflow of the aPC approach, a numerical model to predict permeability evolution of a Lower Tuscaloosa sandstone core exposed to CO2 saturated brine was used. The modeled sandstone core permeability by the aPC approach was 2095.5 mD ± 504.5 mD after 180 days of CO2 exposure. The measured permeability of the core after 180 days of CO2 exposure was 1925.0 mD, which was within the uncertainty range. Keq (SiO2 (am)) was the most important modeling parameter that influenced permeability results, implying that SiO2 (am) is a key mineral that governs permeability evolution of sandstone in geologic carbon storage scenarios. The aPC approach can reduce 99% of simulation time needed to do global sensitivity analysis of a complicated geochemical model, compared with traditional Monte Carlo approach.