A Fiber-sensor-based Long-distance Safety Monitoring System for buried Oil Pipeline

In both industry and research domain, the consensus is that alarm in advance is more valuable than informing after pipeline having been destroyed, and a long-distance monitoring system is more valuable than one with a shorter distance. However, systems reported so far can hardly achieve the requirement. In this paper, we propose a general long-distance (ges50 Km) monitoring system based on distributed optical fiber sensing system. This system uses fiber sensor to collect vibration signal of the soil around the pipeline and then classify different kind activities, including several damaging activities and non-threatening activities, with artificial neural networks. During processing, our system concentrates on balancing between real-time requirement and satisfactory accuracy rate. Experiments show that, compared with other systems, ours is able to alarm in time, and identify activities with an accuracy rate more than 96%, much better than other results reported ever since.