Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
After more than thirty years of continuous research and development, the performances of conventional Brillouin based distributed optical fiber sensors (DOFS) are now peaking, as they are facing fundamental barriers that restrict the power of the light injected into the fiber. For scientists, absolute limitations of this kind are hardly ever accepted as an insurmountable obstacle, rather they are considered as an additional challenge that can always be overcome by a cleverer, and often more complex apparatus. It is thus not surprising to find in the scientific literature a large diversity of reports that describe refinements brought to the conventional architectures of Brillouin based DOFS. In this thesis, we explore two entirely different ways of pushing further the capabilities of conventional Brillouin based DOFS, which are first thoroughly reviewed in a preliminary chapter devoted to depict the most fundamental aspects of this technology.
An entire chapter is dedicated to an interdisciplinary study where the formalism and methods initially developed within the theory of digital signal processing to analyze linear time-invariant systems are transposed to the case of DOFS. This methodology enables, first, to fully understand the potential improvements and limitations in terms of performances of data post-processing when applied to conventional Brillouin optical time-domain analyzers (BOTDA). Then, the concept of deconvolution is presented as a promising tool to achieve sub-metric spatial resolution measurement, which is challenging using direct methods due to the finite lifetime of acoustic phonons in the fiber. Finally, this theory is put at use to revisit the concept of optical pulse coding in BOTDA, demonstrating significant performances improvement over conventional architectures, while circumventing the many practical implementation restrictions met by other coding schemes.
The last section of this dissertation is dedicated to the study of a phenomenon known as forward stimulated Brillouin scattering (FSBS). FSBS is foreseen as a potential candidate to diversify the physical measurands that can be interrogated via DOFS, as it is highly sensitive to the acoustic boundary conditions at the glass outer boundary of the fiber. Far from having the longevity of other branches of DOFS, the performances of recently reported distributed FSBS sensing schemes are still quite poor, hence there is still a large margin for improvement. Here, the acoustic vibrations involved in FSBS are activated harmonically, and the resulting refractive index modulation is picked up by an optical pulse that acts as a pump in a conventional Brillouin based sensor. First, a modified Brillouin optical time-domain reflectometer is implemented, demonstrating distributed FSBS sensing with a SR of 8~m. While displaying interesting features, this method is flawed by severe limitations, notably due to its intensity-based operating principle. We then report, for the first time to the best of our knowledge, a frequency based distributed FSBS sensing technique, that relies on the principle of serrodyne modulation. The results obtained outperform any previously documented reports, achieving a SR of 80~cm in a short section of bare single-mode fiber, and a SR of 2~m over 500~m of polyimide coated fiber.
, , ,
Luc Thévenaz, Tiago Filipe Pimentel Das Neves