Distributed Optical Fiber Sparse Broadband Vibration Signal Sensing Technology Based On Compressed Sensing

Distributed optical fiber sensing technology has more and more applications in the field of health and safety monitoring of large-scale facilities.Among them,?-OTDR technology relies on its real-time sensing and processing ability of vibration signals.It is suitable for the detection of facilities with long-distance,large-scale and real-time characteristics.In order to solve many problems in practical application of vibration sensing,distributed vibration sensing technology is developing towards improving signal-to-noise ratio,extending sensing distance,signal processing parallelism,and widening frequency response range.Increasing frequency response range and accurately locating vibration are helpful to improve the system's ability of feature analysis and discrimination of vibration events.Combined with compressed sensing,the distributed optical fiber sensing system realizes the broadband sparse vibration sensor of distributed system,and realizes the two-dimensional surface location of vibration events in this paper.The main research contents are as follows:The development status of distributed broadband vibration sensing technology and compressed sensing technology is summarized.Combined with the application status of compressed sensing in distributed sensing and analog information converter technology,the feasibility of applying compressed sensing to improve sampling technology in distributed system is analyzed.Two basic theories of distributed broadband vibration based on compressed sensing,distributed vibration sensing based on Rayleigh scattering and compressed sensing technology,and coherent detection technology of distributed system are introduced respectively,which provide guidance for subsequent reconstruction algorithm research.A sparse broadband signal reconstruction method based on matching pursuit is proposed,and the vibration sensing simulation experiment is carried out.The characteristics of sensor matrix under additive random sampling are analyzed,and an improved matching pursuit algorithm is designed.The algorithm uses frequency refinement,least squares solution and time domain fitting techniques to achieve two objectives of spectrum estimation and time domain reconstruction.According to the residual characteristics,the optimal iteration process enhances the recognition range of frequency components.Through simulation experiments,time domain reconstruction of sparse frequency domain signals is realized.A distributed optical fiber vibration sensor system based on additive random interval non-linear frequency modulation pulse is built,and a wide-band sparse vibration sensing experiment is carried out.While using compressed sensing technology to realize sparse broadband vibration signal sensing,according to matched filtering principle,the laser pulse is non-linearly frequency modulated to improve the sensing distance of vibration signal.The sparse vibration signal with frequency range of 20 kHz is successfully restored in time domain on 50 km optical fiber.Vibration positioning experiments are carried out in the experimental tank of winding optical fibers using the engineering prototype of the multi-parameter detection system developed by the laboratory.Firstly,the sparse sampling of the temperature and strain field signals of the experimental tank wall is carried out,and the interpolation algorithm is used to reconstruct the two-dimensional field.Combined with twodimensional interpolation algorithm and edge detection method,the vibration signals are accurately located by analyzing the starting time of each point on the tank wall.