Long-wavelength propagation in fractured rock masses (3D Stress Field)

Abstract

Fractured rocks affect a wide range of natural processes and engineering systems. In most cases, the seismic characterization of fractured rock masses in the field involves wavelengths much longer than the fracture spacing; reproducing this condition in the laboratory is experimentally challenging. This experimental investigation explores the effect of fracture rock fabric and the 3D stress field on P wave propagation in the long‐wavelength regime using a large‐scale true triaxial device. P wave velocities increase with stress in the propagation direction and follow a power law of the form Vp = α(σ’/kPa)β; analyses and experimental results show that stress‐sensitive fracture stiffness and fracture density define the α‐factor and β‐exponent; conversely, long‐wavelength velocity versus stress data can be analyzed to identify the stress‐dependent fracture stiffness. P wave velocities exhibit hysteretic behavior caused by inelastic fracture deformation and fabric changes. During deviatoric loading, the P wave velocity decreases in the two constant‐stress directions due to the development of internal force chains and the ensuing three‐dimensional deformation. Following a load increment, time‐dependent contact deformations result in P wave velocity changes during the first several hours for the tested carbonate rocks; the asymptotic change in velocity is more pronounced for higher stress changes and stress levels. The fracture network geometry that defines the rock fabric acts as a low‐pass filter to wave propagation, so that wavelengths must be longer than two times the fracture spacing to propagate (Brillouin dispersion); the long‐wavelength velocity and the fracture spacing determine the cutoff frequency. Fabric anisotropy contributes to anisotropic low‐pass filtering effects in the rock mass.Plain Language SummarySeismic waves provide a convenient method to characterize fractured rock masses for various applications, from infrastructure engineering to reservoir characterization and production monitoring. But what do wave propagation parameters tell us about the rock mass? Results from this study show that the propagation velocity is a function of the confining stress and that both fracture geometry and stress anisotropy cause velocity anisotropy. We also found that only waves with a wavelength much longer than the spacing between fractures can traverse the medium (short wavelengths get trapped bouncing between fractures); therefore, a high‐frequency cutoff can be used to infer the fracture density in the direction of propagation.Key PointsLong‐wavelength P wave propagation velocity in fractured rocks increases with effective stress due to the stress‐dependent fracture stiffnessRelated phenomena include velocity‐stress hysteresis, fabric‐dependent low‐pass filtering, and creep‐induced stiffness changes in short time scalesLong‐wavelength P wave propagation parameters reflect both fabric and stress anisotropy