Diurnal Changes in Signal-to-Noise Ratio in a Distributed Acoustic Sensing System

Distributed Acoustic Sensing (DAS) systems, typically consisting of a fiber-optic cable and an optical time-domain reflectometry interrogator, are commonly used to detect vibration in the medium surrounding the fiber-optic cable. DAS has been used for over a decade as a low-cost sensor for infrastructure monitoring in oil and gas pipelines. For confidence in DAS as an infrastructure monitoring system, it is important to recognize signal changes with time due to environmental effects. Previous research has explored the long-term changes in system performance. The purpose of this research is to document changes in DAS performance, with signal-to-noise ratio (SNR) in decibels (dBs) as the performance metric, for five data collections throughout each of two days approximately one month apart. One day was dry, and the other day was wet, having experienced over 3 in. (80 mm) of rainfall. The DAS system is installed in a trench containing gravel, sand, flowable fill, and native loess sections. Moisture and temperature sensors in each fill material report hourly average volumetric water content and soil temperature, while a nearby weather station collects hourly rainfall and air temperature data. A calibrated hammer on a metal plate was used to generate seismic waves directly above portions of the array in each fill material, thereby inducing vibrational strains in the fiber-optic cable. These signals were then compared to the noise in the associated cable section immediately before the hammer strike sequence. Results show significant changes in SNR values throughout the day. Consistent with previous research, the gravel trench had the strongest SNR, and the flowable fill had the weakest. While the test trenches were fully saturated on both study days, SNRs on the dry day (a day without precipitation) averaged several dBs higher than on the wet day (a day with precipitation). Variations in SNR within a material over the course of the day were significantly greater on the wet day than the dry day, but the variations within the day do not seem to be correlated to temperature or active rainfall. Understanding diurnal variations in DAS performance will inform the structural health monitoring community to better interpret system results.