16:00 〜 16:15
[STT42-09] 実大三次元震動破壊実験施設を用いたDASの強震動応答性の検討
キーワード：DAS、震動台、強震動、カップリング
Earthquakes significantly impact railway operations, ranging from temporary service restrictions to severe damages, including derailments. In urban areas, even moderate-magnitude earthquakes can cause prolonged service suspensions due to time-consuming post-earthquake inspections (e.g., the 2018 Northern Osaka Earthquake, M6.1; the 2021 Northwestern Chiba Earthquake, M5.9). For railway operations, early detection of seismic motion immediately after an earthquake and rapid assessment of ground shaking along railway lines (for structural damage estimation) are crucial.
Distributed Acoustic Sensing (DAS) measures strain changes along optical fiber cables by analyzing phase variations in backscattered laser pulses. This technology enables high-density strong motion monitoring along railway lines, utilizing existing optical fiber infrastructure. This study evaluates the response characteristics of DAS to strong ground motion using full-scale shaking experiments at the E-Defense facility of the National Research Institute for Earth Science and Disaster Resilience (NIED).
E-Defense is a large-scale shaking table capable of applying three-dimensional seismic motions equivalent to the 1995 Hyogoken-Nanbu Earthquake to full-scale structures (e.g., residential houses, mid-rise buildings). The shaking table measures 20 m × 15 m and can apply ±900 gal in the horizontal directions and ±1500 gal in the vertical direction. In this study, five rows of U-shaped concrete gutters were installed on the shaking table, inside which two types of optical fiber cables were placed: a 200-core telecommunication fiber (Cable A) and a fiber for strain measurement (Cable B). These cables were installed with five different coupling conditions: (1) fully fixed, (2) non-fixed, (3) partially fixed, (4) buried in sand, and (5) suspended along the wall.
The applied seismic waveforms included sine waves with amplitudes ranging from 20 to 1500 gal and frequencies between 0.2 and 8 Hz, as well as recorded seismic waveforms from past earthquakes observed by K-NET. The input seismic motions were applied in three components, while DAS measurements focused on strain changes along the fiber axis. Two DAS interrogators, from AP Sensing and Sintela, were used, with varying parameters during the tests. Additionally, MEMS accelerometers were installed on the shaking table to compare acceleration waveforms with DAS-measured strain rates.
The DAS successfully recorded strain-rate waveforms with predominant frequencies corresponding to the input shaking waveforms. When the optical fiber was rigidly attached to the shaking table and moved as a solid body, no strain was observed. However, in this study, apparent strain changes were detected, likely due to microscopic strain variations in the shaking table material or heterogeneity in the coupling between the fiber and the table.
For Cable A, when fully fixed, strain rate was accurately recorded up to approximately 1000 gal for sine waves below 2 Hz. The strain rate values exhibited a frequency-dependent response relative to the input acceleration. In contrast, the non-fixed optical fiber moved at accelerations above 400 gal, preventing accurate strain rate recording. Furthermore, even for the same input acceleration and frequency, strain rate values varied significantly depending on the coupling condition of Cable A. Weakly coupled sections failed to follow the strain changes of the concrete gutters, leading to discrepancies.
In contrast, Cable B exhibited minimal differences in strain rate across different coupling conditions. This is attributed to its low rigidity and lightweight properties, allowing it to follow strain changes in the gutters with high sensitivity. For the same input sine wave acceleration, the strain rate recorded by Cable B was approximately 10 times higher than that of Cable A. As a result, cycle skipping occurred around 400 gal for all coupling conditions in Cable B.
Distributed Acoustic Sensing (DAS) measures strain changes along optical fiber cables by analyzing phase variations in backscattered laser pulses. This technology enables high-density strong motion monitoring along railway lines, utilizing existing optical fiber infrastructure. This study evaluates the response characteristics of DAS to strong ground motion using full-scale shaking experiments at the E-Defense facility of the National Research Institute for Earth Science and Disaster Resilience (NIED).
E-Defense is a large-scale shaking table capable of applying three-dimensional seismic motions equivalent to the 1995 Hyogoken-Nanbu Earthquake to full-scale structures (e.g., residential houses, mid-rise buildings). The shaking table measures 20 m × 15 m and can apply ±900 gal in the horizontal directions and ±1500 gal in the vertical direction. In this study, five rows of U-shaped concrete gutters were installed on the shaking table, inside which two types of optical fiber cables were placed: a 200-core telecommunication fiber (Cable A) and a fiber for strain measurement (Cable B). These cables were installed with five different coupling conditions: (1) fully fixed, (2) non-fixed, (3) partially fixed, (4) buried in sand, and (5) suspended along the wall.
The applied seismic waveforms included sine waves with amplitudes ranging from 20 to 1500 gal and frequencies between 0.2 and 8 Hz, as well as recorded seismic waveforms from past earthquakes observed by K-NET. The input seismic motions were applied in three components, while DAS measurements focused on strain changes along the fiber axis. Two DAS interrogators, from AP Sensing and Sintela, were used, with varying parameters during the tests. Additionally, MEMS accelerometers were installed on the shaking table to compare acceleration waveforms with DAS-measured strain rates.
The DAS successfully recorded strain-rate waveforms with predominant frequencies corresponding to the input shaking waveforms. When the optical fiber was rigidly attached to the shaking table and moved as a solid body, no strain was observed. However, in this study, apparent strain changes were detected, likely due to microscopic strain variations in the shaking table material or heterogeneity in the coupling between the fiber and the table.
For Cable A, when fully fixed, strain rate was accurately recorded up to approximately 1000 gal for sine waves below 2 Hz. The strain rate values exhibited a frequency-dependent response relative to the input acceleration. In contrast, the non-fixed optical fiber moved at accelerations above 400 gal, preventing accurate strain rate recording. Furthermore, even for the same input acceleration and frequency, strain rate values varied significantly depending on the coupling condition of Cable A. Weakly coupled sections failed to follow the strain changes of the concrete gutters, leading to discrepancies.
In contrast, Cable B exhibited minimal differences in strain rate across different coupling conditions. This is attributed to its low rigidity and lightweight properties, allowing it to follow strain changes in the gutters with high sensitivity. For the same input sine wave acceleration, the strain rate recorded by Cable B was approximately 10 times higher than that of Cable A. As a result, cycle skipping occurred around 400 gal for all coupling conditions in Cable B.