Research On Layered Vibration Extraction And Precision Positioning Algorithm For Distributed Optical Fiber Vibration Sensing System

The distributed optical fiber vibration sensing(DOFVS)system has been widely used in the fields of perimeter security,pipeline security,and communication fiber fault detection due to its characteristics of anti-electromagnetic interference,such as high sensitivity,large spatial range and high positioning accuracy.This paper studies the signal processing technology based on the DOFVS terminal and engineering demonstration system independently developed by the research group.Firstly,in this paper,based on the DOFVS system experimental platform,a series of vibration sensing signal test data is obtained;the time,frequency domain and information entropy characteristics of these signals are deeply studied,and characteristic parameters such as addition / multiplication short-time over-level rate,frequency band standard deviation,segment information entropy,and multiplication integral segment information entropy are proposed and defined separately,find the internal relationship between each feature parameter,which lays the foundation for the system to accurately determine whether there is external vibration.Then,an adaptive threshold filtering algorithm based on the forward translation of Grubbs-standard deviation is proposed,detect vibration at the bottom of the system and extract vibration perception signals,reduce the system false alarm rate and initially improve the system positioning accuracy,verification of the proposed vibration sensing signal extraction algorithm.Through the system to extract the vibration's frequency,vibration's amplitude,vibration's sensing signals of different positions,obtained the response relationship between the characteristic parameters of the vibration sensing signal and the amplitude,frequency and position of vibration.Further,the vibration pulse characteristics of the DOFVS system are analyzed in the upper layer of the system,propose a precise positioning algorithm based on K nearest neighbor(KNN)is proposed,analyze and compare the spatial distribution of vibration positioning results before and after the algorithm,find that they all follow a normal distribution,find a way to locate accurately.The positioning accuracy of the system is improved from ± 258)to ± 2.58)by algorithm,which is 10 times higher;the standard deviation of the vibration positioning's results was reduced from 6m to 0.8m,and the positioning dispersion was reduced by 86%.Finally,the algorithm proposed in this paper is used in the DOFVS project demonstration system,carried out on-site test testing and experimental verification,analyzed and optimized system performance.The results show that the system can accurately locate the single-point vibration position of the full-sensing optical cable in real time,The signal-to-noise ratio(SNR)is 60 d B,the positioning accuracy is ± 2.58),and the response time is 0.6s.