11:30 AM - 11:45 AM
[STT36-09] Development of a Ground Monitoring Method Using Optical Fiber DAS and Microtremor Survey (No.2)
Keywords:optical fiber, Distributed Acoustic Sensing, microtremor survey, ground
Using ground motion measurement data obtained from optical fiber DAS at a test site in Tsukuba city, we applied the extended spatial autocorrelation method and seismic interferometry to determine surface wave phase velocities and estimate a two-dimensional ground structure model. When compared with the results of microtremor exploration conducted around the optical fiber, the phase velocity data showed overall consistency.
Furthermore, starting from September 2023, we conducted ground motion measurements using DAS along National Route 6 and Route 50. The measurement periods for each route were 13 and 10 days, respectively. We set the gauge length to 9.6 meters and the channel spacing to 4.8 meters. The measurement distances were 56.7 km along National Route 6 and 62.8 km along Route 50. The measurement instruments were installed at the Tsuchiura and Iwase branch offices of the National Highways. We conducted seismic interferometry analysis using approximately 30 minutes of data from time 0 hours for about 10 days. Comparing the phase velocities obtained from this analysis with those from past microtremor exploration along National Route 6 and Route 50, we obtained consistent results except for some observation points. For the observation points where consistent results were not obtained, we plan to conduct microtremor exploration to investigate the causes.
We have developed an analysis method combining DAS with seismic interferometry techniques. Specifically, we applied the common midpoint stacking method commonly used in surface wave exploration to the seismic interferometry data recorded by DAS. The observed data is partitioned into segments of 50 channels each. Utilizing waveforms from a 30-minute period starting at midnight each day, we compute the cross-correlation functions between each pair of traces. These cross-correlation functions are then stacked for each measurement day. Furthermore, for all conceivable pairs of traces in the cross-correlation function data, we calculate the cross-correlation functions, gathering those where the midpoint of the two traces is the same. For those with the same receiver spacing, they are co-stacked. Traces with midpoint positions within 25 meters are collected and arranged according to the receiver spacing. By employing the common midpoint stacking method, we were able to distinctly confirm the groups and phase velocities of surface waves propagating from virtual sources, thereby verifying the noise reduction effect. Application of this method allowed us to estimate the 2D S-wave velocity structure directly beneath the fiber.
We developed a prototype system aimed at achieving long-term continuous and real-time monitoring in DAS measurements. The prototype consists of edge processing, which operates on-site to perform noise reduction and downsampling, and storage processing, which operates within the data center to accumulate and retrieve data. We conducted a validation experiment using data collected from National Route 6 and Route 50, inputting the data into the prototype at actual measurement intervals. We used a desktop PC (HP EliteDesk 800 G4 SFF) with standard performance expectations, anticipating installation at the measurement site. To process data in real-time, it is necessary to complete the input-to-output process of edge processing within the length of the time-series data. Through experiments using a file length of 60 seconds, we confirmed the ability to execute edge processing within the data length. As the output of the interrogator is in file format, the processing unit depends on the file, potentially introducing delays. We consider requirements for the data specifications to enhance real-time performance.
Furthermore, starting from September 2023, we conducted ground motion measurements using DAS along National Route 6 and Route 50. The measurement periods for each route were 13 and 10 days, respectively. We set the gauge length to 9.6 meters and the channel spacing to 4.8 meters. The measurement distances were 56.7 km along National Route 6 and 62.8 km along Route 50. The measurement instruments were installed at the Tsuchiura and Iwase branch offices of the National Highways. We conducted seismic interferometry analysis using approximately 30 minutes of data from time 0 hours for about 10 days. Comparing the phase velocities obtained from this analysis with those from past microtremor exploration along National Route 6 and Route 50, we obtained consistent results except for some observation points. For the observation points where consistent results were not obtained, we plan to conduct microtremor exploration to investigate the causes.
We have developed an analysis method combining DAS with seismic interferometry techniques. Specifically, we applied the common midpoint stacking method commonly used in surface wave exploration to the seismic interferometry data recorded by DAS. The observed data is partitioned into segments of 50 channels each. Utilizing waveforms from a 30-minute period starting at midnight each day, we compute the cross-correlation functions between each pair of traces. These cross-correlation functions are then stacked for each measurement day. Furthermore, for all conceivable pairs of traces in the cross-correlation function data, we calculate the cross-correlation functions, gathering those where the midpoint of the two traces is the same. For those with the same receiver spacing, they are co-stacked. Traces with midpoint positions within 25 meters are collected and arranged according to the receiver spacing. By employing the common midpoint stacking method, we were able to distinctly confirm the groups and phase velocities of surface waves propagating from virtual sources, thereby verifying the noise reduction effect. Application of this method allowed us to estimate the 2D S-wave velocity structure directly beneath the fiber.
We developed a prototype system aimed at achieving long-term continuous and real-time monitoring in DAS measurements. The prototype consists of edge processing, which operates on-site to perform noise reduction and downsampling, and storage processing, which operates within the data center to accumulate and retrieve data. We conducted a validation experiment using data collected from National Route 6 and Route 50, inputting the data into the prototype at actual measurement intervals. We used a desktop PC (HP EliteDesk 800 G4 SFF) with standard performance expectations, anticipating installation at the measurement site. To process data in real-time, it is necessary to complete the input-to-output process of edge processing within the length of the time-series data. Through experiments using a file length of 60 seconds, we confirmed the ability to execute edge processing within the data length. As the output of the interrogator is in file format, the processing unit depends on the file, potentially introducing delays. We consider requirements for the data specifications to enhance real-time performance.