Induced Seismicity at Fenton Hill and an Investigation of the Influence of Fault Heterogeneity on the Gutenberg-Richter b-value for Rate-and-State Earthquake Simulations

摘要

In this work, we developed a numerical modeling framework for application to subsurface reservoir engineering problems related to fluid flow, heat transfer, and mechanical deformation in fractured porous media. An existing reservoir simulator was extended to include poroelastic and thermoelastic effects. We introduced a novel approach to couple poroelastic and thermoelastic stresses with fracture deformation calculations. In addition, we implemented a rate-and-state friction model capable of describing the earthquake rupture process on two-dimensional heterogeneous fault surfaces. A spatial random field model is described that is able to generate heterogeneous, spatially correlated, fractal distributions of fault properties such as stress, friction, and permeability. We applied the numerical model to investigate the relative impacts of fracture pressurization, poroelastic stress, and thermal stress on the evolution of seismicity observed during reservoir stimulation treatments at the Fenton Hill enhanced geothermal system test site. A geologic conceptual model of the site was developed based on several independent datasets. We then simulated several scenarios to isolate the roles of each physical mechanism that could potentially contribute to seismicity. Although the magnitudes of the poroelastic and thermal stresses were more than an order of magnitude lower than the change in fluid pressure in the fractures during injection, they contributed significantly to the seismicity rate and migration pattern. Comparing the simulation results with the actual earthquake catalog data suggests that poroelastic and thermoelastic effects likely did play a role during the Fenton Hill stimulation treatments. In another study, we hypothesized that by introducing heterogeneous and spatially correlated patterns of fault properties, it would be possible to simulate earthquake sequences that exhibit power law scaling of frequency-magnitude distributions using a rate-and-state friction model. We examined the influence of heterogeneous distributions of initial shear stress, initial normal stress, dynamic friction coefficient, and permeability. Our model was able to simulate earthquake sequences exhibiting Gutenberg-Richter-type frequency-magnitude distributions with b-values close to 1 when the fault had heterogeneous stress or friction. Permeability heterogeneity did not have an effect. This study has implications for understanding how the Gutenberg-Richter b-value is influenced by natural geologic conditions and injection well operations.