5:15 PM - 6:45 PM
[STT36-P06] Estimation of 2D strain tensors using DAS data around fiber bending points
Keywords:Distributed Acoustic Sensing, Strain tensor, fiber bending points
Distributed acoustic sensing (DAS) is a technique to measure strain (or strain rate) time series at dense spatial points along fiber optic cables. Only a single component of strain tensors, axial strain along optical fibers, can be obtained. For subsurface imaging using surface waves, straight sections of the fibers are usually used. Fiber optic cables buried under loads or banks usually have some bending points. At these bending points, the polarity can change for SH and Love waves (e.g., Martin et al., 2021). Based on these polarity characteristics, Zhao et al. (2022) succeeded in estimating the energy ratio between Rayleigh and Love waves. Therefore, DAS data at these bending points are helpful for some purposes. Here, we propose a method to estimate 2D strain tensors from DAS data around bending points of an optical fiber and explore the method's applicability using DAS data at Sakurajima volcano in Japan.
We assume that DAS channels are located on a flat surface by neglecting the topography. X axis is taken along East-West direction, and y axis is taken along North-South direction. According to the transformation of strain tensor by the coordinate rotation, the DAS axial strain at one channel is expressed with areal strain (exx+eyy), differential extension (exx-eyy) and the shear strain exy. Assuming the spatial homogeneity of the strain field, we are able to set up an inverse problem to estimate three unknowns (areal strain, differential extension, and shear strain around the fiber bending points) from DAS data at more than three channels.
Among the DAS data obtained during our observation at Sakurajima from November to December 2022, we used data along a traverse line encircling Sakurajima. The length is approximately 39km, channel separation is 4.8m, gauge length is 9.6m, and the sampling frequency is 200Hz. Among the channels, we selected three bending points around channels 570 (Higashi-Sakurajima), 1530 (Arimura), and 2640 (Kurokami) according to the following reasons. The route is bending, the topography change is not large, and strain seismograms look similar. We also selected an earthquake of Mw6.1 that occurred at Off Mie Prefecture because such a moderate-size earthquake can generate long-period waves. Record sections show coherent signals with a predominant period of 3-5 seconds. We also confirm that the polarity of strain records is sometimes reversed at these bending points. Taking five channels with a step of 5 channels (ca. 24m) at both sides of a bending point, we estimate areal strain, differential extension, and shear strain by solving the inverse problem.
Around channel 570, shear strain is dominant, and its amplitude is 7x10^-8 strain at maximum. We confirmed that polarity reversals are found on the record section when the shear strain becomes large. At JMA Yuno seismic station, located approximately 500m to the East, seismograms show the predominance of the horizontal components. Particle motions suggest that SH or Love waves come from the East. That can explain the observed large shear strains. Around channel 1530, DE is the largest, with an amplitude of 1.3x10^-7 strain. That is different from strains around channel 570. These two channels are approximately 2km apart, which is similar to the seismic wavelength with periods of 3-5 s. Around channel 2640, a car passed, causing noise. The car noise is characterized by the large DE. At JMA Seto station, which is located approximately 200m to the north, seismograms are dominated by the horizontal components. However, the particle motion with the horizontal plain was complicated. This is different from the characteristics found for the JMA Yuno station.
The proposed method assumes the spatial homogeneity of strain fields. Therefore, the gauge length should be much smaller than the seismic wavelength. Site amplification effects and the coupling between fibers and the ground should be homogeneous as well. Moreover, the accurate geo-localization of channels (azimuth) is also essential. Although there are such constraints, we can extract information about the 2D strain tensor from data near the bending points of the optical fiber if we choose appropriate locations.
Acknowledgments
We used fiber optic cables from the Ministry of Land, Infrastructure, Transport and Tourism.
We thank the Osumi River National Highway Office for helping us with the DAS observation. We used data at volcanic observation stations of the Japan Meteorological Agency.
We assume that DAS channels are located on a flat surface by neglecting the topography. X axis is taken along East-West direction, and y axis is taken along North-South direction. According to the transformation of strain tensor by the coordinate rotation, the DAS axial strain at one channel is expressed with areal strain (exx+eyy), differential extension (exx-eyy) and the shear strain exy. Assuming the spatial homogeneity of the strain field, we are able to set up an inverse problem to estimate three unknowns (areal strain, differential extension, and shear strain around the fiber bending points) from DAS data at more than three channels.
Among the DAS data obtained during our observation at Sakurajima from November to December 2022, we used data along a traverse line encircling Sakurajima. The length is approximately 39km, channel separation is 4.8m, gauge length is 9.6m, and the sampling frequency is 200Hz. Among the channels, we selected three bending points around channels 570 (Higashi-Sakurajima), 1530 (Arimura), and 2640 (Kurokami) according to the following reasons. The route is bending, the topography change is not large, and strain seismograms look similar. We also selected an earthquake of Mw6.1 that occurred at Off Mie Prefecture because such a moderate-size earthquake can generate long-period waves. Record sections show coherent signals with a predominant period of 3-5 seconds. We also confirm that the polarity of strain records is sometimes reversed at these bending points. Taking five channels with a step of 5 channels (ca. 24m) at both sides of a bending point, we estimate areal strain, differential extension, and shear strain by solving the inverse problem.
Around channel 570, shear strain is dominant, and its amplitude is 7x10^-8 strain at maximum. We confirmed that polarity reversals are found on the record section when the shear strain becomes large. At JMA Yuno seismic station, located approximately 500m to the East, seismograms show the predominance of the horizontal components. Particle motions suggest that SH or Love waves come from the East. That can explain the observed large shear strains. Around channel 1530, DE is the largest, with an amplitude of 1.3x10^-7 strain. That is different from strains around channel 570. These two channels are approximately 2km apart, which is similar to the seismic wavelength with periods of 3-5 s. Around channel 2640, a car passed, causing noise. The car noise is characterized by the large DE. At JMA Seto station, which is located approximately 200m to the north, seismograms are dominated by the horizontal components. However, the particle motion with the horizontal plain was complicated. This is different from the characteristics found for the JMA Yuno station.
The proposed method assumes the spatial homogeneity of strain fields. Therefore, the gauge length should be much smaller than the seismic wavelength. Site amplification effects and the coupling between fibers and the ground should be homogeneous as well. Moreover, the accurate geo-localization of channels (azimuth) is also essential. Although there are such constraints, we can extract information about the 2D strain tensor from data near the bending points of the optical fiber if we choose appropriate locations.
Acknowledgments
We used fiber optic cables from the Ministry of Land, Infrastructure, Transport and Tourism.
We thank the Osumi River National Highway Office for helping us with the DAS observation. We used data at volcanic observation stations of the Japan Meteorological Agency.