Seismic monitoring of fractures

My dissertation research focused on the seismic signature of reservoir rock fractures, experimentally informing the detectability and monitoring of large-scale fluid injection, such as in many energy applications including geothermal, carbon storage, and hydro-fracking. Unique in rock physics, I modified a seismic-frequency (0.1-100 Hz) apparatus for measuring shear anelasticity (modulus and attenuation) under a range of conditions (reactive fluid flow, strain amplitude, and uniaxial confining stress). Combining the fractures’ anelasticity with images of normal stress distribution from pressure sensitive film, we modeled the shear mechanics of asperity contacts as well as estimate the seismic detectability of fractures for the range of applicable effective stresses [Saltiel et al., Geophysics 2017]. At the most extreme conditions of low normal and high shear stresses, we found strong nonlinear elasticity associated with friction on slipping contact areas. This data was consistent with a simple analytic partial slip model, allowing inversion of the effective contact area and friction coefficient. By extrapolating the model to full sliding, this study suggests controlled high amplitude seismic surveys could be used to estimate frictional properties and predict the conditions for shear failure in situ [Saltiel et al., GRL 2017]. Since these measurements are made using oscillating shear stresses over a range of frequencies, slip is induced at a constantly changing velocity and the data contain information of the rate dependence as well as transient frictional behavior. The shape of nonlinear hysteretic curves is dominated by the effect of static friction during velocity reversal. Since common fault friction laws are not designed for zero or negative velocities, a friction model developed by robotics engineers (bristle-state) was needed to predict the response to oscillations. Inversions of a simple spring-slider block model with this parameterization show that these sort of seismic data could also provide constraints on dynamic friction properties, those that determine if faults fail in earthquakes, slow slip, or creep [Saltiel et al., JGR 2017]. 

Collaborators: Jonathan Ajo-Franklin (Rice); Brian Bonner (LBNL); Paul Selvadurai (ETH); Tushar Mittal (Penn State); Brent Delbridge (Los Alamos)