Integration of reservoir simulation, history matching, and 4D seismic for CO2-EOR and storage at Cranfield, Mississippi, USA

In this paper, we compare 4D seismic interpretations of CO2 plume evolution with fluid-flow numerical simulation results for Cranfield, Mississippi. Historic pressure trends, oil and gas production rates, and current CO2-EOR production data from the field were history matched, and a tuned model was used for predictive simulations. For CO2-EOR operations, numerical simulation results of the CO2 plume distribution and CO2 first arrival (breakthrough) times in production wells were compared to the available field data. Three interpretations of 4D seismic data show discrepancies on the edges of the seismic survey, and along the sealing fault, where numerical simulations show high CO2 saturations. In areas between these two limits, the match between simulation and 4D seismic interpretation improves. In addition, for most of the production wells, comparison of the breakthrough time of CO2 showed a reasonable match. The tuned model was then used to predict reservoir response and storage capacity in different field development scenarios under CO2 injection. We compared hypothetical scenarios where the operator transitions from CO2-EOR to CO2 injection without oil production (CO2-EORT) when oil production is not economical anymore, to a scenario of continuing with CO2-EOR. Our results show that CO2-EOR can store more CO2 and operations will last longer, whereas if switched to CO2-EORT, the field must be abandoned earlier because of spillover of the CO2 plume. However, the amount of CO2 stored per year is larger for CO2-EORT as compared to CO2-EOR. (C) 2016 Elsevier Ltd. All rights reserved.