Research On Vibration Signal Processing And Reconstruction Algorithm Based On φ-OTDR System

The distributed fiber optic sensing is attracting people to pay more attention.Compared with other traditional electronic sensors,distributed fiber optic sensing has passive,anticaustic and flexible advantages.It plays an important role in pipeline leakage,power monitoring and mine exploration.Among them,because of the distributed fiber optic sensing based on φ-OTDR technology has advanges of long detecting distance,high spatial resolution and wide frequency response,φ-OTDR is always research hotspot.At present,on the one hand,the φ-OTDR system misjudges location of the vibration due to system noise and surroundings noise.On the other hand,in traditional φ-OTDR system,limited by Nyquist sampling theorem,the biggest measurement frequency of vibration signal is only half of the pulse repetition frequency.So the frequency measurement range of(p-OTDR system is becoming reduce with the sensing distance increased.This thesis offered solutions of total variation algotithm and compressed sensing technology based on previous research for impact of noise on system performance and contradictory relation between frequency response range and pulse repetition frequency.And the signal to noise of the system is improved and the high frequency signal can be reconstructed at low pulse repetition frequency.The main research and work of this thesis are as follows:(1)The distributed fiber sensing is introduced in this thesis in the first place and the research background and significance are introduced.Via classification of the distributed fiber optic sensing,the advantages and disadvantages of various structural system are pointed out.Secondly,the research progress at home and abroad of φ-OTDR system in recent years are introduced and classified according to performance parameter.Then,the basic principles of traditional OTDR and cp-OTDR systems are introduced in detail.In the end,the main parameters for evaluating the performance of φ-OTDR system are discussed(2)This thesis proposed a solution of signal processing,which of φ-OTDR system based on total variation algorithm.It.realized the promotion of system performance by reduce background noise.Firstly,the backward Rayleigh scattered light signals of φ-OTDR system are transformed into gray images.According to Euler-Lagrange multiplier constraint and gradient descent method,the denosing process was implemented.The number of iterations was optimized according to experimental result to improve SNR of the system.Secondly,compared with the common moving average moving difference algorithm and the non-local mean algorithm,the advantages of total variation in the process of cp-OTDR system denoising are shown.Then,the location experiment were carried out.It used total variation algorithm to locate single vibration and multi vibration and verified the applicability and accuracy.The experiment results prove that the vibration event of 300 Hz can be located accurately at the length of 3.5 km fiber,and the SNR reaches 10.13 dB.Finally,the recognition ability of the total variation algorithm to broadband signal and the actual vibration signal was tasted,and the broadband signal of 300 Hz to 1 kHz was successfully identified,and the actual vibration position was accurately located(3)The compressed sensing technology is applied to the φ-OTDR system to solve the contradiction between the pulse repetition frequency and the frequency response range,and realize the measurement and reconstruction of the high-frequency vibration signal sampled by sub-Nyquist.Firstly,the thesis introduces the principle of compressed sensing technology.Then using the constructed the sub-Nyquist sampling random sequence pulse modulation mode respectively for units of single frequency of vibration signal,the unit of the multifrequency vibration signal and multi position multifrequency vibration signal disturbance position detection and signal reconstruction,on average less than 2 kHz pulse repetition frequency under the condition of complete reconstruction of 8 kHz vibration signal.The main innovations of this thesis are as follows:(1)The φ-OTDR system is studied from the perspective of two-dimensional image.The total variation algorithm is used to deal with the noise of the system,convert the backward-scattering Rayleigh light signal into gray image,and use the Euler-Lagrange multiplier constraint and gradient descent method to reduce the noise of the φ-OTDR system,and improve the performance of the φ-OTDR system.The proposed total variation algorithm provides a new scheme for signal processing in cp-OTDR system.(2)A random monopulse modulation scheme is proposed based on compressed sensing technology.Under the condition of the pulse repetition frequency which is far lower than Nyquist’s sampling theorem,the high frequency vibration signal in the φ-OTDR system is reconstructed,and the frequency response range of the system is improved.(3)The sub-Nyquist random monopulse sequence sampling is used to effectively reduce the pressure of data storage and signal processing in the(p-OTDR system,which lays an important foundation for the research of the long-distance high-frequency distributed fiber sensing system.