Research On Optical Fiber Vibration Monitoring Based Φ-OTDR

As a new kind of distributed fiber optic sensor,phase sensitive optical time domain reflectometry has incomparable advantages in remote and real-time monitoring of vibration signal in a wide range and long distance compared to conventional sensors.At this stage,it is the focus of international competition in the field of hi-tech security.Φ-OTDR is a kind of distributed sensor based on backward Rayleigh scattering. It needs extremely narrow linewidth and low frequency drift laser as the light source to construct the system. Inject the periodic optical pulse into the fiber, detect the dynamic vibration signal around the fiber, locate the alarm point according to the light pulse round-trip time between the detection place and the light source. The linewidth of light source is about 10kHz, so, the signal received by detector is the interence results of backward Rayleigh scattering within the light pulse width. External vibration will pruduce fiber index modulation, resulting in minor fluctuations of fiber index, which cause fluctuation of interference results, it is the foundation of vibration monitoring using Φ-OTDR.In this thesis, based on summarizing previous works, completed the following works:(1) Details the Φ-OTDR system, it includes basic principles, system architecture, performance indicators and relations of mutual restraint between these indicators.(2) Combines the captured signal and the characteristic of device used by the system, analyses the characteristic of backward Rayleigh scattering signals. Put forward the wavelet denoising solution and it do well in recurring the optical power curve in case of external vibration. Generating the hot spot figure of the system, and do threshold processing, the target point is point greater than threshold, realizing the accurate judgement of alarm point in the distributed system by doing clustering analysis for target points.(3) After finished the signal denoising, alarm point judgement, obtain the feature vectors of the vibration information fragments, get the feature vectors of 8 dimensions. The collected samples were divided into two parts, training and testing, design pattern recognition classifier based on artificial neural network, achieved a good recognition result.