2:30 PM - 2:45 PM
[SSS07-09] Evolution of a slip surface properties revealed by transmitted elastic waves during rotary shear friction experiments
Keywords:transmitted elastic wave, gouge layer, friction coefficient
Information on fault friction properties is essential to mitigate disasters from induced seismicity. Because it is impossible to directly observe the friction surface where shear slip occurs, transmitted elastic waves were used to measure the condition of the friction surface in the lab experiments. Fukuyama et al. (2017) conducted rotary shear friction experiments using transmitted elastic waves through the slip surface to investigate the relationship between the friction coefficient and temporal changes in the wave amplitude.Fukuyama et al. (2017) experimentally confirmed the correlation between the maximum amplitude of transmitted elastic waves through a simulated fault and the friction coefficient. They suggested that these amplitude changes were due to changes in the size of the voids within the gouge layer. Based on this, we computed the waveforms propagating through cracks distributed on the sliding surfaces to imitate a simulated gouge layer in this research. Elastic waves were calculated by the wavefield calculation code OpenSWPC (Maeda et al. (2017)) using the finite difference method to reproduce transmitted elastic waves. By comparing the computed waveforms with the observed ones by Fukuyama et al. (2017), we aim to estimate the heterogeneous structure of the gouge layer in the fault plane during shear slip.
First, to calibrate the piezoelectric sensor response, an elastic wave transmission test was conducted using a cylindrical specimen of metagabbro with a diameter of 38 mm and a height of 82.5 mm. The same piezoelectric elements as in Fukuyama et al. (2017) were attached to the upper and lower surfaces of the specimen, and half-cosine waves of 50 V amplitude and 500 kHz frequency were input to the piezoelectric elements on the upper surface at 0.005 s intervals using a function generator with an amplifier. The elastic waves generated by the piezoelectric element propagate through the specimen and were recorded as a voltage signal by the piezoelectric element on the bottom surface.
In the finite difference simulation, the sample was represented as a cylinder with a diameter of 38 mm and a height of 82.5 mm with a grid interval of 0.1 mm, which was created within a cuboid space with 481x481x826 grids. The time step was set to 5.0×10-9 s. Inside the cylinder, we assume the density of 2.980 kg/m3, the P-wave velocity of 5729 m/s, and S-wave velocity of 3308 m/s. A half-cosine source time function with a period of 2.0×10-6s was applied perpendicular to the surface at the center of the upper surface of the cylinder, and elastic waves were computed at the center of the lower surface. The response function of the piezoelectric sensor was obtained by Fourier transforming the voltage of the elastic wave measured by the piezoelectric sensor and the velocity amplitude waveform of the elastic wave obtained by the finite difference calculation, respectively.
Referring to the cylinder specimen, the numerical sample model was constructed as a cylinder with a diameter of 25.05 mm, a height of 70 mm, and a grid interval of 0.1 mm, in a cuboid space of 351 x 351 x 700 grids. The time step was calculated at an interval of 5.0×10-9s for 10,000 steps. For the cylindrical part, the physical properties (density 2.980 kg/m3, P-wave velocity 6919 m/s, S-wave velocity 3631 m/s) of Fukuyama et al. (2017) were used, and outside the cylinder the medium was assumed to be a vacuum. A half-cosine source time function with a period of 2.0×10-6s was input perpendicular to the surface at the center of the top surface of the cylinder. Cracks representing the gouge layer were set up as pseudo-voids with P-wave velocity of 1000 m/s, S-wave velocity of 500 m/s, and density 0.001kg/m3. An ensemble of cracks was located perpendicular to the direction of wave propagation at 34.5 mm from the top surface. And we changed the crack density within the layer as a controlling parameter. The calculated waveform multiplied by the response function of the piezoelectric element was compared with the waveform obtained by Fukuyama et al. (2017) to estimate the time variation of the crack density parameter.
First, to calibrate the piezoelectric sensor response, an elastic wave transmission test was conducted using a cylindrical specimen of metagabbro with a diameter of 38 mm and a height of 82.5 mm. The same piezoelectric elements as in Fukuyama et al. (2017) were attached to the upper and lower surfaces of the specimen, and half-cosine waves of 50 V amplitude and 500 kHz frequency were input to the piezoelectric elements on the upper surface at 0.005 s intervals using a function generator with an amplifier. The elastic waves generated by the piezoelectric element propagate through the specimen and were recorded as a voltage signal by the piezoelectric element on the bottom surface.
In the finite difference simulation, the sample was represented as a cylinder with a diameter of 38 mm and a height of 82.5 mm with a grid interval of 0.1 mm, which was created within a cuboid space with 481x481x826 grids. The time step was set to 5.0×10-9 s. Inside the cylinder, we assume the density of 2.980 kg/m3, the P-wave velocity of 5729 m/s, and S-wave velocity of 3308 m/s. A half-cosine source time function with a period of 2.0×10-6s was applied perpendicular to the surface at the center of the upper surface of the cylinder, and elastic waves were computed at the center of the lower surface. The response function of the piezoelectric sensor was obtained by Fourier transforming the voltage of the elastic wave measured by the piezoelectric sensor and the velocity amplitude waveform of the elastic wave obtained by the finite difference calculation, respectively.
Referring to the cylinder specimen, the numerical sample model was constructed as a cylinder with a diameter of 25.05 mm, a height of 70 mm, and a grid interval of 0.1 mm, in a cuboid space of 351 x 351 x 700 grids. The time step was calculated at an interval of 5.0×10-9s for 10,000 steps. For the cylindrical part, the physical properties (density 2.980 kg/m3, P-wave velocity 6919 m/s, S-wave velocity 3631 m/s) of Fukuyama et al. (2017) were used, and outside the cylinder the medium was assumed to be a vacuum. A half-cosine source time function with a period of 2.0×10-6s was input perpendicular to the surface at the center of the top surface of the cylinder. Cracks representing the gouge layer were set up as pseudo-voids with P-wave velocity of 1000 m/s, S-wave velocity of 500 m/s, and density 0.001kg/m3. An ensemble of cracks was located perpendicular to the direction of wave propagation at 34.5 mm from the top surface. And we changed the crack density within the layer as a controlling parameter. The calculated waveform multiplied by the response function of the piezoelectric element was compared with the waveform obtained by Fukuyama et al. (2017) to estimate the time variation of the crack density parameter.