Optical Fiber Sensor for Real-Time Monitoring of Industrial Structures and Application to Urban Telecommunication Networks.

Distributed optical fiber sensors are important for continuous remote monitoring of large infrastructures, such as gas and oil pipelines, civil controlled perimeters, dams, roads, railroads, and also telecommunication networks. We present the theoretical study and practical implementation of a phase-sensitive distributed fiber sensor, capable of real-time monitoring of an urban area telecommunication network. The traditional data processing is here enhanced by CNN machine learning algorithms providing recognition accuracy >98 % for different perturbation or intrusion processes in the inspection wells located along the fiber network. In addition to this specific application to a telecommunication network, other applications of this fiber sensor are possible -and also already implemented- for gas&oil pipelines, railroads, power plants, and transportation and industrial sites in general.