Signal-to-Noise Analysis Of Computational Brillouin Optical Time-Domain Analysis

With the rapid development of modern sensing technology,Brillouin optical time-domain analysis(BOTDA)which can realize online real-time monitoring,gradually becomes one of the major long-distance distributed fiber-optic sensing(DFOS)technologies.The system has many advantages,such as long-distance measurement capability,high spatial resolution and high measurement precision.It has been widely used in monitoring the health of some massive structures,such as railway system,bridges,dams,oil and gas pipelines,and so on.However,compared to the widely used traditional sensors,the BOTDA system requires relatively longer measurement time and higher cost,which greatly limits its application scope.In order to reduce the system's cost,the computational BOTDA technique is proposed based on the differential ghost imaging principle which can significantly reduce the sampling rate requirement of the data acquisition module compared to that in the conventional BOTDA system.Over 3 order of magnitude reduction can be achieved experimentally.This thesis mainly studies the basic principle of the computational BOTDA technique,and analyze its signal-to-noise ratio in detail.The main contents of this thesis are as follows:Firstly,the advantages,classifications,and the basic working principle of BOTDA are introduced.In addition,the fundamental principle of BOTDA and its development are presented.The basic principle of Walsh-Hadamard transform,ghost imaging,classical ghost imaging,and differential ghost imaging are also introduced.Secondly,the linear relationship between the Brillouin frequency shift with respect to the temperature and strain variations is analyzed theoretically.It is also proved that the differential ghost imaging protocol using the binary sequence pattern from a Walsh-Hadamard matrix can be used to reconstruct the time-domain Brillouin scattering signals perfectly.At the same time,MATLAB software is used to simulate the reconstruction theory and verify the reconstruction equation.Then,based on the theoretical analysis and simulation results,a traditional BOTDA system and a computational BOTDA system based on the differential ghost imaging protocol are built in the laboratory,and the key components used in the computational BOTDA system are introduced.As a demonstration purpose,time-domain traces of a 1 km sensing fiber with a1 m spatial resolution are successfully reconstructed.Finally,the influence of the order of the Walsh-Hadamard matrix and the number ofaverages on the noise of the computational BOTDA system is theoretically analyzed based on the establishment of the mathematical model,simulation verification and a large number of collected experimental data.Both simulation results and experimental results show that when only the white Gaussian noise is considered in the detection,the computational method requires twice more number of averages compared to the conventional time-domain method to achieve the same signal-to-noise level.Since the computational approach is focusing on stationary measurement,doubling the measurement time can normally be acceptable in practice,but it can reduce the sampling rate requirement significantly compared to the conventional method,offering great advantage to simplify the data acquisition design in the distributed fiber-optic sensing system.