COMPARATIVE ASSESSMENT OF CO2-SEQUESTRATION IN THE FARNSWORTH, TEXAS HYDROCARBON FIELD USING THE STOMP-EOR, TOUGHREACT, AND GEM REACTIVE TRANSPORT SIMULATORS

The objective of the present study was to investigate the injection and storage of CO₂ in the Morrow B Sandstone, the principal reservoir in the Farnsworth, Texas hydrocarbon field, using three numerical reactive transport simulators, STOMP-EOR, TOUGHREACT, and GEM. The models in this investigation were based on a five-spot well pattern centered on one of the main injection wells in the Farnsworth field. Two model scenarios were investigated: (1) A “saline aquifer” aquifer scenario in which water was the only initial pore fluid present; (2) an “enhanced oil recovery (EOR)” scenario in which water and oil were the initial pore fluids present. The saline aquifer scenario was modeled with all three simulators. The enhanced oil recovery scenario was modeled only with STOMP-EOR and GEM. In both model scenarios, CO₂ was injected on a water-alternating-gas schedule for 25 years, after which injection ceased and the simulations were allowed to continue for another 975 years to track reservoir behavior. Within each model scenario, model inputs were the same to facilitate comparison of the simulators’ performances.

Some consistent results were observed in the model results. Temperature decreased and pressure increased during the injection phase, after which the trends reversed and pressure and temperature homogenized over time. All of the models predicted a sharp decrease in pH near the injection well, which inhibited carbonate mineral precipitation. All of the models predicted calcite to dissolve continuously and for dolomite to be a major mineral sink for CO₂. All of the models predicted small changes in porosity over time, but too small to impact the permeability and fluid flow patterns significantly. The models differed in other respects. The models predicted significant differences in silicate mineral abundance. In the EOR scenario, oil was consistently predicted to be the main sink for injected CO₂. In the saline aquifer model, either immiscible CO₂ or aqueous solution could be the main sinks for injected CO₂. The results of the study not only provide insights into the behavior of injected CO₂ but also uncertainties in this behavior based on the degree of consistency in the outputs of the different numerical simulators.