9:30 AM - 9:45 AM
[SSS04-03] Exploring the application of Characteristic Functions on DAS data and their influence in event detection performance.
Keywords:Distributed Acoustic Sensing, DAS, detector, Characteristic Functions
Distributed Acoustic Sensing (DAS) has become a significant asset in seismology, repurposing the fiber-optic cables into dense arrays of geophones that can continuously sample the seismic wavefield over several kilometers. The versatility of DAS data acquisition allows for flexible deployment methods, including ad hoc installations with dedicated cables or tapping into existing telecommunication optical fiber networks. This adaptability proves especially beneficial for seismic monitoring in demanding environments such as volcanoes or offshore regions, where traditional seismometers encounter logistical constraints.
Traditional seismological methods face challenges in efficiently handling Distributed Acoustic Sensing (DAS) data, primarily due to its distinctive attributes. Indeed, seismometers are sparse installations having sampling frequency generally of 250 Hz or less, producing a small amount of data. On the other hand, DAS usually sample seismic wavefields at 1m spacing and with frequencies that may exceed 1kHz, which provide a detailed mapping of the wavefield over the all length of the fiber, but also produce terabytes of data per day. The requirement of new algorithms to analyze these data lead to the development of a waveform-based detection method specifically designed for DAS data (Porras et al. 2024).
In this work, we apply several characteristic functions (CFs) to DAS data in order to enhance signal and mitigate noise. These functions are both non-negative and zero-mean. The first ones evaluated are the Short Term Average to Long Term Average (STALTA), Energy and Envelope, whose characteristic is to preserve noise. On the other hand, the second ones mitigate noise. Examples are the Short Term Average to Long Term Average derivative (STALTA derivative), the Kurtosis and the Kurtosis derivative. Our target is to evaluate our detector performance in analyzing preprocessed data compared to raw ones.
Our tests are at first carried out on synthetic data, simulated using different optical fiber geometries, source configuration and fiber locations. In a second stage we apply the detector on real data collected in two different scenarios. The initial scenario centers around the FORGE experiment located in Utah, USA, where a borehole is cabled with a 1 km optical fiber placed above a geothermal reservoir known for induced seismic events. The second scenario involves a 90 km horizontal optical fiber in the Pyrenees region, renowned for both natural earthquakes with magnitudes ranging from 0.02 to 2.01 and human-induced seismic events due to quarry blasts.
Our assessment is centered on measuring the improvement in the detector performance after applying Characteristic Functions (CFs) compared to analyzing the raw data. Through this extensive investigation, our objective is to advance the comprehension of DAS data processing while showcasing the detector’s effectiveness across a spectrum of scenarios.
We would like to thank TotalEnergies for sharing the Pyrenees data set with us as well as Febus Optic for providing the DAS interrogator used for the data acquisition.
References:
A Semblance-based Microseismic Event Detector for DAS Data.
J. Porras, D. Pecci, G. Bocchini, S. Gaviano, M. De Solda, K. Tuinstra, F. Lanza, A. Tognarelli, E. Stucchi, F. Grigoli. Geophysical Journal International (GJI) 2024
Traditional seismological methods face challenges in efficiently handling Distributed Acoustic Sensing (DAS) data, primarily due to its distinctive attributes. Indeed, seismometers are sparse installations having sampling frequency generally of 250 Hz or less, producing a small amount of data. On the other hand, DAS usually sample seismic wavefields at 1m spacing and with frequencies that may exceed 1kHz, which provide a detailed mapping of the wavefield over the all length of the fiber, but also produce terabytes of data per day. The requirement of new algorithms to analyze these data lead to the development of a waveform-based detection method specifically designed for DAS data (Porras et al. 2024).
In this work, we apply several characteristic functions (CFs) to DAS data in order to enhance signal and mitigate noise. These functions are both non-negative and zero-mean. The first ones evaluated are the Short Term Average to Long Term Average (STALTA), Energy and Envelope, whose characteristic is to preserve noise. On the other hand, the second ones mitigate noise. Examples are the Short Term Average to Long Term Average derivative (STALTA derivative), the Kurtosis and the Kurtosis derivative. Our target is to evaluate our detector performance in analyzing preprocessed data compared to raw ones.
Our tests are at first carried out on synthetic data, simulated using different optical fiber geometries, source configuration and fiber locations. In a second stage we apply the detector on real data collected in two different scenarios. The initial scenario centers around the FORGE experiment located in Utah, USA, where a borehole is cabled with a 1 km optical fiber placed above a geothermal reservoir known for induced seismic events. The second scenario involves a 90 km horizontal optical fiber in the Pyrenees region, renowned for both natural earthquakes with magnitudes ranging from 0.02 to 2.01 and human-induced seismic events due to quarry blasts.
Our assessment is centered on measuring the improvement in the detector performance after applying Characteristic Functions (CFs) compared to analyzing the raw data. Through this extensive investigation, our objective is to advance the comprehension of DAS data processing while showcasing the detector’s effectiveness across a spectrum of scenarios.
We would like to thank TotalEnergies for sharing the Pyrenees data set with us as well as Febus Optic for providing the DAS interrogator used for the data acquisition.
References:
A Semblance-based Microseismic Event Detector for DAS Data.
J. Porras, D. Pecci, G. Bocchini, S. Gaviano, M. De Solda, K. Tuinstra, F. Lanza, A. Tognarelli, E. Stucchi, F. Grigoli. Geophysical Journal International (GJI) 2024