RESUMO
This study compares noise and signal-to-noise ratio (SNR) in direct detection and coherent detection fiber-based distributed acoustic sensing (DAS) systems. Both detection schemes employ the dynamic analysis of Rayleigh-backscattered light in phase-sensitive optical time-domain reflectometry (ΦOTDR) systems. Through theoretical and experimental analysis, it is determined that for photodetection filters with a sufficiently narrow bandwidth, the SNR performance of both detection schemes is comparable. However, for filters with poor selectivity, coherent detection was found to exhibit superior performance. These findings provide crucial guidelines for the design of high-performance time-domain DAS systems.
RESUMO
Distributed acoustic sensors (DAS) perform distributed and dynamic strain or temperature change measurements by comparing a measured time-domain trace with a previous fiber reference state. Large strain or temperature fluctuations or laser frequency noise impose the need to update such a reference, making it necessary to integrate the short-term variation measurements if absolute strain or temperature variations are to be obtained. This has the drawback of introducing a 1/f noise component, as noise is integrated with each cumulative variation measurement, which is detrimental to the determination of very slow processes (i.e., in the mHz frequency range or below). This work analyzes the long-term stability of chirped-pulse phase-sensitive optical time-domain reflectometry (CP-ΦOTDR) with multi-frequency database demodulation (MFDD) to carry out "calibrated" measurements in a DAS along an unmodified SMF. It is shown that, under the conditions studied in this work, a "calibrated" chirped-pulse DAS (CP-DAS) with a completely suppressed reference update-induced 1/f noise component is achieved capable of making measurements over periods of more than 2 months with the same set of references, even when switching off the interrogator during the measurement.