Esprit Estimation Algorithm For Central Frequency Of Scattering Spectrum In Brillouin Optical Time-Domain Reflectrometry

Based on Brillouin Optical Time Domain Reflectometry(BOTDR)distributed optical fiber sensing system,temperature and strain information along long-distance optical fiber along the path can be measured in real time,with high spatial resolution,accuracy.It has gradually become a mainstream technology in the field of fiber optic sensing,and is widely used in electricity,petroleum,mineral exploration,structural health monitoring and other fields.The content of this thesis includes the following aspects:First,the basic theory of distributed sensing technology based on BOTDR is introduced,and the generation mechanism of stimulated Brillouin scattering and spontaneous Brillouin scattering is briefly described and compared.The center frequency of Brillouin scattering spectrum signal is studied.The response characteristics and quantitative relationship of temperature changes are described in detail.The main performance parameters of the distributed optical fiber sensing system based on BOTDR are described in detail.Second,the Brillouin scattering spectrum model is described,and the classification of spectrum estimation methods is briefly introduced.Because in BOTDR system,the estimation of the center frequency of the Brillouin scattering spectrum at various places in a fiber is important to measurement,and it is also the most time-consuming part,it is quite difficult to make BOTDR system quickly responsed.Therefore,this thesis proposes to use ESPRIT algorithm to estimate Brillouin scattering spectrum center frequency in BOTDR system,which is compared with the frequency estimation algorithm based on Fast Fourier Transform(FFT).The results show that ESPRIT algorithm has performance similar to that obtained by Lorentz spectrum fitting with a zero-padded FFT algorithm at a lower request for computational resources.Thirdly,a temperature sensing experiment system based on BOTDR is built.Firstly,the data collected at different temperatures are estimated by ESPRIT algorithm and FFT algorithm respectively,and the reliability of ESPRIT algorithm is verified.Second,with various reduced data length,the average of standard deviations of the frequency measurements of the two algorithms in multiple measurements is compared,verifying the stability of the ESPRIT algorithm,Then the average of the absolute difference between the measured temperature and the actual temperature of each algorithm under different data lengths is calculated to evaluate the accuracy of ESPRIT algorithm,Finally,the average absolute temperature difference of ESPRIT algorithm at different pulse widths is analyzed to evaluate its frequency estimation performance at the highest spatial resolution.The results show that ESPRIT algorithm requires less data length and can obtain better frequency estimation performance.Therefore,it can improve the measurement speed with higher spatial resolution and measurement performance.