The integrity of a geological formation is a primary concern in any underground fluid injection project. Hydraulic pressurization due to injection may reduce fault strength, trigger fault slippage, and cause fault reactivation. The reactivated fault affects the fluid migration and loss from the injection zone, which might undermine the efficiency and safety of the project. Hence, a reliable modeling of fault reactivation is critical.
In this work, we propose a new approach to modeling fault reactivation. Faults are complex structures and generally consist of core and damage zones with macroscopic fracture networks. The embedded discrete fracture model (EDFM) is an effective method for simulating complex geometries such as fracture networks and nonplanar hydraulic fractures. We used the EDFM in conjunction with a compositional reservoir simulator to model fault reactivation under hydraulic pressurization. The phase behavior and fluid flow are accurately modeled using the equation of state (EOS) compositional simulation.
The activation of fault occurs at a threshold pressure, which depends on the chemo-mechanical properties of the formation rock. The threshold pressure can be estimated using analytical, numerical, or laboratory methods. In this study, we provided an analytical calculation of the threshold pressure. Moreover, we used a refined, multiphase, compositional, and geomechanical reservoir simulator to predict the threshold pressure. The coupled geomechanical reservoir simulation is computationally expensive; therefore, we suggest using this approach, in the absence of laboratory measurements, to simulate only a few regions of the formation with distinctive rock types. The estimated values of threshold pressures for different geomechanical rock types can be used in our simulations.
We performed large-scale reservoir simulations using the EDFM to investigate the storage capacity of carbon depositional formations representative of the Gulf of Mexico and monitor CO2 migration paths before and after fault reactivation. The results of this study are helpful to evaluate the capacity and integrity of carbon storage sites. Our methodology gives promising results for the prediction of fault reactivation and CO2 migration within a formation.
The proposed approach accurately models faults and their reactivation. It does not require refinement and geomechanical calculation for each gridblock in the domain, which reduces the computational time by at least five times. The significance of this approach becomes more pronounced in large formations with multiple rock types and faults. Although we used our approach for the study of carbon storage, the same methodology can be used for other types of fluid injection, such as water disposal.