Geochemical modeling of CO2 storage in deep reservoirs: The Weyburn Project (Canada) case study

Abstract

Geological storage is presently one of the most promising options for reducing anthropogenic emissions of CO2. Among the several projects investigating the fate of CO2 stored at depth, the EnCana's CO2 injection EOR (Enhancing Oil Recovery) project at Weyburn (Saskatchewan, Canada) is the most important oil production development that hosts an international monitoring project. In the Weyburn EOR Project CO2 is used to increase recovery of heavy oil from the Midale Beds, a Mississippian reservoir consisting of shallow marine carbonate, where about 3 billions standard m3 of supercritical CO2 have been injected since 2000 with an injection rate of 5000 ton/day. In this work the available dataset (bulk mineralogy of the reservoir, gas-cap composition and selected preand post-CO2 injection water samples) provided by the International Energy Agency Weyburn CO2 Monitoring & Storage Project has been used in order to: i) reconstruct the pre-injection reservoir chemical composition (including pH and the boundary conditions at 62 °C and 15 MPa); ii) assess the evolution of the reservoir subjected to CO2 injection and predict dissolution/precipitation processes of the Weyburn brines over 100 years after injection; iii) validate the short-term (September 2000–2003) evolution of the in situ reservoir fluids due to the CO2 injection, by comparing the surface analytical data with the composition of the computed depressurized brines. To achieve these goals the PRHEEQC (V2.14) Software Package was used with both modified thermodynamic database and correction for supercritical CO2 fugacity. The oil–gas–water interaction and the non-ideality of the gas phase (with exception of CO2) were not considered in the numerical simulations. Despite intrinsic limitations and uncertainties of geochemical modeling, the main results can be summarized, as follows: 1) the calculated pre-injection chemical composition of the Midale Beds brine is consistent with the analytical data of the waters collected in 2000 (baseline survey), 2) the main reservoir reactions (CO2 and carbonate dissolution) take place within the first year of simulation, 3) the temporal evolution of the chemical features of the fluids in the Weyburn reservoir suggests that CO2 can safely be stored by solubility (as CO2(aq)) and mineral trapping (via dawsonite precipitation). The short-term validation performed by calculating chemical composition of the reservoir fluids (corrected for surface conditions) after the simulation of 3 years of CO2 injection is consistent (error ≤5%) with the analytical data of the wellhead water samples collected in 2003, with the exception of Ca and Mg (error N90%), likely due to complexation effect of carboxilic acid.