Distributed optical fiber sensing based on frequency-scanned phase-sensitive optical time-domain reflectometry

Unlike traditional sensing schemes which rely on discrete point sensors that perform measurements at predetermined positions, distributed optical fiber sensing is a general technique that enables to continuously gather information (typically temperature and/or strain) along the entire length of an optical fiber. This remarkable feature fulfills today's high demand for multi-points monitoring and detection in complex industrial facilities, e.g. gas/oil delivery systems and civil infrastructures. As one of the most recently investigated distributed optical fiber sensors, frequency-scanned phase-sensitive optical time-domain reflectometry ($\phi$-OTDR) can provide quantitative measurements with high sensitivity by frequency shift estimation using the local Rayleigh backscattering spectra, showing the potential of long-distance and high-spatial-resolution sensing. In this thesis, particular attention has been paid to this technique to achieve more reliable measurements and explore diverse applications rather than conventional temperature and strain sensing.

After a general introduction on distributed optical fiber sensing, including the main scattering phenomena occurring in optical fibers and the most common interrogation methods, the working principle and numerical model, as well as the phase demodulation schemes of $\phi$-OTDR, are reviewed. An emphasis is placed on frequency-scanned $\phi$-OTDR, especially in terms of temperature, strain, and pressure sensitivity when using standard single mode fibers and some other sensing performance criteria.

A thorough study on the issue of large-errors when using cross correlation for the frequency shift estimation is given analytically and experimentally. The results yielding an explanation to the problem and providing with an estimation of the large error probability. In a disruptive approach, a least mean square algorithm is proposed to reduce these errors, enabling to experimentally demonstrate temperature sensing with long distance and high spatial resolution. Distributed pressure sensing, distributed and dynamic strain sensing, as well as distributed hydrogen sensing, are finally presented and demonstrated using different fibers, based on frequency-scanned $\phi$-OTDR with high spatial resolution (5~cm/20~cm). The sensing principles, experimental setups, and results are presented in full detail. The study demonstrates the capability of frequency-scanned $\phi$-OTDR for different applications going beyond temperature and static strain sensing, in particular when combined with fibers based on dedicated designs.