How Does Injection Rate Effect Failure Threshold in Fluid-Filled Fault Gouge?

The question of how earthquakes may be triggered by fluids has become increasingly important in recent decades, driven largely by observations that link subsurface fluid injections to subsequent nearby earthquakes. It is clear that understanding and predicting fluid-triggering is required in many energy-related activities (e.g. geothermal energy, CO2 injection), as well as for naturally triggered events. One of the main open questions is the effect of fluid-injection rate. Usually the 'effective stress law' is invoked to predict the failure of fluid-saturated granular or porous media. This law assumes that in fluid-pressurized faults the instantaneous value of pore pressure controls fault strength and failure. But recent laboratory results (Passelègue et. al., 2018) suggest that the level of pressure by itself cannot describe the full mechanics. Experiments and field work show that the rate of fluid injection is also important: slower injections lead to failure at lower pressures than fast injection rates.

We shall present results from a coupled hydromechanical-discrete element model that simulates the response of a pre-stressed fault, filled with a granular fault gouge, to fluid injection at different rates. Our simulation results find similar rate dependence as seen in the laboratory experiment, i.e. that slow injection causes failure at lower fluid pressure than faster injection. We explore and contrast two end member mechanisms to explain these observations: pore-pressure heterogeneity and fault creep.