The civil engineering structures become, on one hand more and more complicated and/or aged and on the other hand, the society seeks to increase their lifetime which require a special attention to prevent the damages: leaks, fires, landslides, earthquakes, etc... Thus to do so, a large network of sensors is necessary, and its insertion inside the structure has often to be planned before its construction. This led to a new field of research and techniques: Structural Health Monitoring (SHM). Thanks of their intrinsic small size, electromagnetic immunity and robustness, optical fibers sensors offered many advantages compared to existing electronic strain sensors (vibrating strings, linear variable differential transformers), or temperature sensors (thermocouples, resistive probes). A lot of point-like optical fiber sensors based on Bragg gratings or Fabry-Perot cavities have been developed in the nineties, and their multiplexing is still an intensive research topic nowadays, mainly on the aim of cutting down the whole sensing system costs. About ten year ago, distributed optical fiber sensing technologies appeared, constituting a breakthrough in sensor technology. Those methods, based on non-linear optical effects (Raman, Brillouin) or light backscattering, allow to measure temperature and strain (and consequently pressure, displacement ) into a high range zone, wherever the user want, with tunable spatial resolution, bringing consequently huge multiplexing possibilities. Now this technology has become quite mature and suitable for industry, thanks to the creation of many companies providing distributed optical fibers sensing devices. Nevertheless, several topics are still intensively studied. Indeed, distributed optical strain sensors are always sensitive to temperature as well. Unfortunately, in some real structures as bridges, temperature and strain can both vary significantly at the same time. We will demonstrate, with the example of a distributed Brillouin sensor located into a reinforced concrete beam, that it is very difficult to distinguish thermal and mechanical effects on measurements, even if the physical origin of the process is well known. Spatial resolution limit of these methods is tightly linked to the chosen technology. Non-linear distributed optical sensors (Raman and Brillouin) are based on optical pulse-echo technique. Then, the resolution is proportional to optical pulse time-width, and limited by the phonon lifetime (10 ns). In these cases the minimum spatial resolution is restricted to approximately 1m, and the range can reach 20km because it is only limited by the optical losses (about 0.2dB/km). For some applications, requiring high resolution (sink-hole detection), the Optical Backscattering Reflectometry (OBR) technique could be a better solution, thanks to its resolution better than 1cm. However, the range is then limited to about 100m, mainly due to the light coherence length. Thus, the sensor choice is always a trade-off between long-range and high-resolution. However, between those extreme cases, despite a few promising laboratory experiments, there is no commercial solution for both middle-range (1km) and middle-resolution (5 cm) requirements at the present time, to the best of our knowledge.
