Research On Signal Processing Technology Of Distributed Optical Fiber Sensing System

Distributed optical fiber sensing techniques have been widely used in many fields such as security,1 military,and pipeline transportation because of the optical fiber's small size,light weight,no power supply requirement,anti-electromagnetic interference.Fiber optic sensing systems based on different fiber loop structures are very different in detection range,signal-noise ratio,sensitivity and so on.Among them,the optical time domain reflection based on phase-sensitive Rayleigh scattering(?-OTDR)has become the hot spot in the field of distributed optical fiber sensors,because of its simple structure,high sensitivity and simultaneous detection and location of multi-point vibration.In this paper,a series of signal processing methods for positioning,noise reduction and classification of vibration signals are proposed on the basis of deeply studying and comparing the principle of?-OTDR fiber vibration sensor.The main research contents are as follows:1.The basic principle of ?-OTDR is introduced,and a distributed optical fiber sensing system is constructed.The experiment results show that the amplitude and frequency of the vibration signal detected by the system is linearly related to the intensity and frequency of the vibration source.The system improvement schemes respectively based on 3×3 demodulation technology and Signac high birefringence fiber loop are proposed.By analyzing the signal-to-noise ratio of vibration signal and its linearity with the intensity of the vibration source,the detection effect of two improved systems are compared with the original system.2.In this paper,time-domain characteristics,frequency-domain characteristics and scale-domain characteristics of vibration signals are studied deeply.Aiming at the characteristics of ?-OTDR system with large amount of real-time data and susceptibility to environmental disturbances,a vibration detection method based on dynamic characteristic thresholds of short-time energy,short-time double-level rate,frequency band variance and energy variance of wavelet coefficients.This method enables accurate real-time detection of vibration events occurring on the sensing fiber.3.Aiming at the problem that the signal-to-noise ratio of the vibration is low in the long-distance monitoring environment and it is difficult to identify the vibration event by single feature parameter,a vibration event recognition method based on spectral subtraction noise reduction-multi feature extraction-support vector machine is proposed.4.A ?-OTDR distributed optical fiber sensor system signal processing platform is built in this paper.The paper analyzes the whole framework and working flow of the system,introduces the work flow of the signal acquisition system and the host computer monitoring software platform and finally describes the test methods and results of the main indicators of the signal processing platform.