2:30 PM - 2:45 PM
[SSS07-04] Investigation of frictional properties of metagabbro gouge in centimeter scale for understanding scale dependence of rock friction
Keywords:Gouge friction, Scale dependence, Rock friction experiment
When a numerical simulation is conducted to reproduce or to predict a fault rupture, frictional properties of rock estimated from centimeter-scale samples are often used as a premise for the calculations, which implicitly assumes the independence of rock friction on the fault scale. However, it is unclear whether the assumption is proper or not. To make it clear, we are conducting the rock friction experiments in various scales. Since natural faults generally contain crushed rock particles called gouge in the core, we report the experimental results conducted with a centimeter-scale fault containing a simulated gouge layer here.
To directly compare the results with those in meter scale reported by Yamashita et al. (2020, SSJ meeting) and Shimoda et al. (2020, SSJ meeting), we also used metagabbro particles ground with a jet mill as the simulated gouge (mean diameter: 12 µm, maximum diameter: 75 µm) and distributed it on the simulated fault so that the initial thickness before the compression was 3 mm. We conducted the experiments by using a biaxial friction apparatus (Mizoguchi et al., 2021, EPS) at CRIEPI. The dimension of the simulated fault was 10 cm in length by 5 cm in width. The normal stress was maintained at 3.4 MPa or 6.7 MPa during the experiments, same as in the above studies in meter scale.
We first applied the normal stress for 10 minutes, and then sheared the gouge layer for 4 mm at the slip velocity of 10 µm/s, which was followed by the velocity step-changes of 0.1-1.0-10.0-100.0 µm/s for three times. The amount of the shear displacement was 0.5 mm at each slip-velocity stage except for 1.0 mm at the slip velocity of 100.0 µm/s. We assumed that the evolution of friction after the velocity step change obeys the slip law (Ruina, 1983, JGR) of the rate- and state-dependent friction law (RSF law hereafter) and estimated the related parameters using a computer program proposed by Skarbek and Savage (2019, Geosphere) for fitting the observed curve of friction.
The estimated RSF parameters, a-b and Dc, show weak dependences on the cumulative slip displacement and the loading rate. The values of estimated a-b are consistent with those estimated in the meter-scale experiments. Dc estimated in this study ranges from several to around 10 µm, which is consistent with the fact that the mean diameter of the simulated gouge is 12 µm. However, in the meter-scale experiments, Dc reached more than 100 µm at maximum, and thus those are much larger than the current ones. We will check if this difference is significant or not, and then investigate the cause of the difference if significant.
To directly compare the results with those in meter scale reported by Yamashita et al. (2020, SSJ meeting) and Shimoda et al. (2020, SSJ meeting), we also used metagabbro particles ground with a jet mill as the simulated gouge (mean diameter: 12 µm, maximum diameter: 75 µm) and distributed it on the simulated fault so that the initial thickness before the compression was 3 mm. We conducted the experiments by using a biaxial friction apparatus (Mizoguchi et al., 2021, EPS) at CRIEPI. The dimension of the simulated fault was 10 cm in length by 5 cm in width. The normal stress was maintained at 3.4 MPa or 6.7 MPa during the experiments, same as in the above studies in meter scale.
We first applied the normal stress for 10 minutes, and then sheared the gouge layer for 4 mm at the slip velocity of 10 µm/s, which was followed by the velocity step-changes of 0.1-1.0-10.0-100.0 µm/s for three times. The amount of the shear displacement was 0.5 mm at each slip-velocity stage except for 1.0 mm at the slip velocity of 100.0 µm/s. We assumed that the evolution of friction after the velocity step change obeys the slip law (Ruina, 1983, JGR) of the rate- and state-dependent friction law (RSF law hereafter) and estimated the related parameters using a computer program proposed by Skarbek and Savage (2019, Geosphere) for fitting the observed curve of friction.
The estimated RSF parameters, a-b and Dc, show weak dependences on the cumulative slip displacement and the loading rate. The values of estimated a-b are consistent with those estimated in the meter-scale experiments. Dc estimated in this study ranges from several to around 10 µm, which is consistent with the fact that the mean diameter of the simulated gouge is 12 µm. However, in the meter-scale experiments, Dc reached more than 100 µm at maximum, and thus those are much larger than the current ones. We will check if this difference is significant or not, and then investigate the cause of the difference if significant.