Fault Stabilization by Dilatant Hardening in Granular Rocks
Taka KanayaPublished August 15, 2019, SCEC Contribution #9881, 2019 SCEC Annual Meeting Poster #165
Triaxial compression experiments were conducted on Fontainebleau sandstone with initial porosities of 4 and 6% at a constant effective pressure of 70 MPa (produced by various combinations of confining and pore pressures), a strain rate of 10-5/s, and room temperature. Our experiments show a transition from unstable to more stable failure with an increase in absolute confining and pore pressures, which is more evident in the low porosity samples. For 4% porosity samples, at Pc = 80 MPa (i.e., Pf = 10 MPa), failure occurs within <1 second with significant acoustic emissions. In contrast, at Pc = 140 to 200 MPa (i.e., Pf = 70 to 130 MPa), 4% porosity samples fail over 10 to 100 sec. with few AEs. For 6% porosity samples, all samples fail within a <1 sec.; failure results in significant AEs at Pc = 80 to 140 MPa, whereas limited AEs are observed at Pc = 200 MPa. Relative to samples showing unstable failure, samples showing stable failure display higher peak strength and bulk modulus during the elastic portion of axial deformation. These observations suggest that the samples deformed at high absolute Pc and Pf are undrained throughout elastic loading and failure, in which the stabilization of failure likely results from dilatant hardening.
Permeabilities were determined for these samples under hydrostatic loading at effective pressures to 70 MPa. The permeability shows a strong dependence on sample porosity (10-19 m2 at 4% porosity to 10-13 m2 at 6% porosity) and a modest dependence on effective pressure.
Using these results, we test the hypothesis that undrain failure is promoted when the time scale of bulk sample diffusion is greater than that of a characteristic strain. For 4% porosity samples, the observed permeability, sample storage, and sample length yield a diffusion time much longer than the entire duration of axial deformation. This suggests that all of our 4% porosity axial deformation tests were undrained, consistent with the observed stable failure. For 6% porosity samples, our analysis yields a diffusion time much shorter than a time scale of axial strain of 0.1 to 1%, but conformable with that of unstable faulting (~1 sec.). The former predicts that all of our 6% porosity samples were drained, not consistent with the limited AEs observed at Pc = 200 MPa. Hence, we conclude that the drainage condition, and thus the stability, of failure are governed by the timescales of bulk diffusion and unstable faulting.
Citation
Kanaya, T. (2019, 08). Fault Stabilization by Dilatant Hardening in Granular Rocks. Poster Presentation at 2019 SCEC Annual Meeting.
Related Projects & Working Groups
Fault and Rupture Mechanics (FARM)