| With the increasing operating mileage of high-speed railways,it greatly facilitates people’s travel while also adding potential hazards to railway safety.The traditional measures for preventing intrusion along railway lines mainly include physical fences,wire mesh,and infrared radiation,which have high installation costs and environmental requirements;Distributed optical fiber sensing technology has the advantage of detecting physical quantities(such as strain and temperature)at any position within the sensing distance,and has attracted much attention in the fields of structural health monitoring,power system and pipeline leakage monitoring.Among them,Phase-sensitive optical time domain reflectometer(Φ-OTDR)has the advantages of high sensitivity and fast response to vibration signal detection,making it very suitable for application in high-speed railway security engineering.This article is based on the Φ-OTDR distributed optical fiber vibration sensing system,research and comparative analysis are conducted on the extraction,noise reduction,and positioning methods of vibration signals;On this basis,we taking the Φ-OTDR system has been applied in the intrusion monitoring project of high-speed railway boundaries,and has conducted monitoring research on simulated pedestrian intrusion events.The specific content is as follows:Firstly,introducing the causes and characteristics of rayleigh scattering effect in optical fibers,and evaluating Optical time-domain reflectometer and Φ-OTDR which are based on rayleigh scattering effect,providing principle explanation,formula derivation,and analysis of technical differences.In addition,introducing the principles and analyze the sensing mechanisms of the two signal detection methods-the distributed vibration sensing and distributed acoustic sensing which are based on Φ-OTDR,and providing a detailed explanation of the main performance indicators of theΦ-OTDR system.Secondly,building the Φ-OTDR distributed optical fiber vibration sensing system,proposing a method for denoising vibration signals using singular spectrum analysis algorithm and designing an experimental to test the system frequency measurement performance.On the basis of moving average and moving difference methods for vibration signal processing,singular spectrum analysis algorithm is used to denoise the vibration signal,which can increase the signal-to-noise ratio by 2.31 d B;Simultaneously conducing experimental measurements on vibration signals of four different frequencies: 40 Hz,100Hz,200 Hz,and 1000 Hz,using algorithms such as fast fourier transform,improved empirical mode decomposition,and short time fourier transform to analyze vibration signals in time and frequency domains,verified the frequency measurement performance of the Φ-OTDR system,demonstrating its potential for application in high-speed rail perimeter intrusion monitoring engineering.Finally,building the high-speed rail perimeter intrusion monitoring system which is based on the Φ-OTDR system,proposing a solution to event pattern recognition which is based on the Φ-OTDR system combines Open CV deep learning algorithm and conducting pedestrian intrusion monitoring research along actual high-speed railways.On the basis of signal normalization,data preprocessing,and cleaning,using singular spectrum analysis algorithm to denoise the spatiotemporal image of the signal and using it as input samples for event pattern recognition,by comparing three typical classifiers: artificial neural network,support vector machine,and K-nearest neighbor algorithm,the recognition rates were 92.2%,87.8%,78.9% respectively.On this basis,artificial neural networks was used as the pattern recognition algorithm for the highspeed rail perimeter intrusion monitoring system in this paper for engineering field experiments.The results showed that the correct recognition rate of the system was95%,which verified the feasibility of applying Φ-OTDR system to high-speed rail perimeter intrusion monitoring. |
PDF Full Text Request