2:15 PM - 2:30 PM
[SSS07-11] Evaluation of the waveform characteristics generated from heterogeneous structures in a rock sample through comparison of observation and simulation
Keywords:rock sample, wave propagation, pore, FDM
Seismic waveforms contain information about underground structures that they pass through. Understanding the relationship between the heterogeneous structure and elastic waveforms will help us monitor the underground structure through the waveforms. Previously, laboratory elastic wave transmission tests were performed to detect the crack generation process in a rock sample (e.g., Yukutake, 1989, JGR). Laboratory studies estimate the heterogeneity change in a sample from the first wave arrival time and amplitude changes, while we obtain little information from later phases. This is because phases repeatedly reflect and refract in a small finite rock sample and generate complicated later phases. Yoshimitsu et al. (2016, GJI) reproduced the elastic waves transmitting through a homogeneous sample with the 3D finite difference method (FDM) and detected the generation process of the characteristic phases observed in the later phases in a transmitted wave. In this study, we update their work and observe the elastic wave propagation process under the heterogeneous sample using FDM simulations.
Matsuda et al. (2023, AGU) conducted elastic wave transmission tests using three rock samples with different heterogeneous characteristics and observed the difference in the transmitted waveforms. Here, we refer to the characteristics of porous andesite sample AO1 collected in Aomori prefecture and build the simulation model. Then, FDM wave propagation simulations are performed with OpenSWPC (Maeda et al., 2017, EPS). We discretize the rectangular calculation area into 1024×1024×2048 with 50 μm intervals, which include a cylindrical sample with 50 mm in radius and 100 mm in height. AO1 includes a variety of pores filled with air that radius ranges from millimeters to several centimeters. We introduce the largest pore to the model with a radius of 30 mm, located 60 mm in height and 90 degrees from the vibration input plane. Referring to Matsuda et al. (2023, AGU), we set the elastic parameters as Vp=4.2, Vs=2.4, rho=2.23. We input a single force for the 50 mm in height. The time step of the calculation is 0.005 sec with the accuracy of the 2nd order in space and time. We also calculate a homogeneous model for the comparison.
We applied a 4th order butterworth bandpass filter between 50 kHz and 1 MHz for the waveforms. The wavefields are calculated from the divergence and rotation of P- and S-wave amplitudes. In the homogeneous model, a body wave is generated at the source and propagates through the sample to the opposite side of the sample. During the propagation, curved boundary generates reflections and refractions. After reaching the opposite side of the sample, the reflected waves propagate back to the source direction. In the pore included model, the propagation path of the body wave swerves from the straight path since the spherical shape of the pore affects the wavefield. The pore also affects reflected waves, and wavefields continue to distort the propagation direction over time. Because of this distortion, the observed amplitude is larger on one side of the sample, and symmetric propagation is broken. In addition, the homogeneous case shows clear surface waves propagating along the circumference edge of the sample, while the pore included case shows a significant disturbance of the waveforms in later phases. Characteristics of the later phases in the inhomogeneous sample show good agreement with the observation by Matsuda et al. (2023, AGU); thus, the size of the pore and position strongly affect the wavefield and shape of the waveforms. The results of this study indicate that the waveforms obtained from array-like multiple stations have a possibility to estimate the location and size of inhomogeneities in the underground.
Matsuda et al. (2023, AGU) conducted elastic wave transmission tests using three rock samples with different heterogeneous characteristics and observed the difference in the transmitted waveforms. Here, we refer to the characteristics of porous andesite sample AO1 collected in Aomori prefecture and build the simulation model. Then, FDM wave propagation simulations are performed with OpenSWPC (Maeda et al., 2017, EPS). We discretize the rectangular calculation area into 1024×1024×2048 with 50 μm intervals, which include a cylindrical sample with 50 mm in radius and 100 mm in height. AO1 includes a variety of pores filled with air that radius ranges from millimeters to several centimeters. We introduce the largest pore to the model with a radius of 30 mm, located 60 mm in height and 90 degrees from the vibration input plane. Referring to Matsuda et al. (2023, AGU), we set the elastic parameters as Vp=4.2, Vs=2.4, rho=2.23. We input a single force for the 50 mm in height. The time step of the calculation is 0.005 sec with the accuracy of the 2nd order in space and time. We also calculate a homogeneous model for the comparison.
We applied a 4th order butterworth bandpass filter between 50 kHz and 1 MHz for the waveforms. The wavefields are calculated from the divergence and rotation of P- and S-wave amplitudes. In the homogeneous model, a body wave is generated at the source and propagates through the sample to the opposite side of the sample. During the propagation, curved boundary generates reflections and refractions. After reaching the opposite side of the sample, the reflected waves propagate back to the source direction. In the pore included model, the propagation path of the body wave swerves from the straight path since the spherical shape of the pore affects the wavefield. The pore also affects reflected waves, and wavefields continue to distort the propagation direction over time. Because of this distortion, the observed amplitude is larger on one side of the sample, and symmetric propagation is broken. In addition, the homogeneous case shows clear surface waves propagating along the circumference edge of the sample, while the pore included case shows a significant disturbance of the waveforms in later phases. Characteristics of the later phases in the inhomogeneous sample show good agreement with the observation by Matsuda et al. (2023, AGU); thus, the size of the pore and position strongly affect the wavefield and shape of the waveforms. The results of this study indicate that the waveforms obtained from array-like multiple stations have a possibility to estimate the location and size of inhomogeneities in the underground.