The influence of hydraulic fracturing in carbon storage performance
Pengcheng Fu
Lawrence Livermore National Laboratory
- Date & Time
- Location
- Building 3, Rambo Auditorium
- Host
- Jack Norbeck
- Summary
Conventional principles of the design and operation of geologic carbon storage (GCS) require injecting CO2 below the caprock fracturing pressure to ensure the integrity of the storage complex. In non-ideal storage reservoirs with relatively low permeability, modest injection rates can lead to pressure buildup and hydraulic fracturing of the reservoir and caprock. While the GCS community has generally viewed hydraulic fractures as a key risk to storage integrity, a carefully-designed stimulation treatment could provide improved injectivity while maintaining overall seal integrity. A vertically-contained hydraulic fracture, either in the reservoir rock or extending a limited height into the caprock, provides an effective means to access reservoir volume far from the injection well. Employing a fully-coupled numerical model of hydraulic fracturing, solid deformation, and fluid flow, we study the likelihood, processes, and consequences of hydraulic fracturing during CO2 injection. A hydraulic fracture’s pressure-limiting behavior dictates that the near-well fluid pressure is only slightly higher than the fracturing pressure of the rock and is insensitive to injection rate and mechanical properties of the formation. Although a fracture contained solely within the reservoir rock, with no caprock penetration, would be an ideal scenario, poroelastic principles dictate that sustaining such a fracture would require continuously increasing pressure until the caprock is fractured. We also investigate the propagation behavior and injection pressure responses of a hydraulic fracture propagating in a caprock subjected to heterogeneous in situ stress. The results could have important implications for the use of geomechanically-contained hydraulic fracturing as a tool to manage storage performance.
The simulations in this study were performed on GEOS, a massively-parallel, fracture-centric, code developed at the Lawrence Livermore National Laboratory for computational geosciences applications. We will provide a brief overview of this code at the seminar.