[SCG53-P21] Stress and Acoustic Emission measurements in granular stick-slip
Keywords:granular matter, stick-slip, acoustic emission, earthquake, tremor
Slow slip and tremor are observed to occur simultaneously (e.g., Rogers and Dragert, Science, 2003). Such observation can be modelled using granular stick-slip experiments which shows that the timing of slips and acoustic emission (AE) are well correlated (e.g., Johnson et al., GRL, 2013, Jiang et al. GRL, 2017). AE waveforms are complex (e.g., Michlmayr & Or, Gran. Matt., 2014), and we do not fully understand the details of the temporal change of AE during the slip. Here we report the results of granular stick-slip experiments which focuses on the relation between the shear stress and AE, and their dependence on the shear rate, particle size and pore fluid.
We shear glass beads using a rotating viscometer (Higashi and Sumita, JGR, 2009), and measure the shear stress and AE, using torsion spring and accelerometers, respectively. Mean shear stress decreases with shear rate, and also when particle size is smaller or when the granular matter is water-immersed. AE acceleration amplitude increases with shear rate and particle size, and decreases for a water-immersed case. AE velocity – time series data, obtained by integrating acceleration, indicates that there are two characteristic frequency (f) components ; a high f (of the order of 100 Hz) and a low f (of the order of 10Hz) components, and that the former tends to end earlier than the latter. Using the bandpass filtered data, we calculate the energy median time for the high f and low f components, and showed that for most cases, h-f component precedes the low-f component. We also confirmed a positive correlation between the AE power and the work per unit time needed for shearing.
We hypothesize that the high-f component corresponds to the particle readjustment time. We extend the“Singing sand” model (Bagnold, 1966) and find that it explains the high-f of the order of 100. In order to evaluate the effect of fluid viscosity, we calculated the Stokes number (St) for the situation in our experiments. We find that St = 41 for the water-immersed case, indicating that inertia dominates over viscosity for particle size motions. This explains why fluid viscosity had small effect on AE in our experiments. The broad-band waveform analyses which we conducted using our experimental data suggest that such analyses may be useful to investigate the details of the relation between slow earthquakes and tremor.
We shear glass beads using a rotating viscometer (Higashi and Sumita, JGR, 2009), and measure the shear stress and AE, using torsion spring and accelerometers, respectively. Mean shear stress decreases with shear rate, and also when particle size is smaller or when the granular matter is water-immersed. AE acceleration amplitude increases with shear rate and particle size, and decreases for a water-immersed case. AE velocity – time series data, obtained by integrating acceleration, indicates that there are two characteristic frequency (f) components ; a high f (of the order of 100 Hz) and a low f (of the order of 10Hz) components, and that the former tends to end earlier than the latter. Using the bandpass filtered data, we calculate the energy median time for the high f and low f components, and showed that for most cases, h-f component precedes the low-f component. We also confirmed a positive correlation between the AE power and the work per unit time needed for shearing.
We hypothesize that the high-f component corresponds to the particle readjustment time. We extend the“Singing sand” model (Bagnold, 1966) and find that it explains the high-f of the order of 100. In order to evaluate the effect of fluid viscosity, we calculated the Stokes number (St) for the situation in our experiments. We find that St = 41 for the water-immersed case, indicating that inertia dominates over viscosity for particle size motions. This explains why fluid viscosity had small effect on AE in our experiments. The broad-band waveform analyses which we conducted using our experimental data suggest that such analyses may be useful to investigate the details of the relation between slow earthquakes and tremor.