Research On Signal Processing Of BOTDA Fiber Sensor System

At the end of the 20th century,the rapid development of optical fiber communication industry showed a very strong practicability in various fields.China has a large land area,and for long-distance communication equipment,transportation equipment and information transmission equipment,it is necessary to timely monitor the health situation,prevent and reduce the occurrence of disasters.According to such requirements,the optical fiber sensor system gradually enters the people's field of vision.Due to the distributed optical fiber sensing system is easy to carry,light weight,the advantages of strong anti-jamming capability,The application of long distance large facilities in monitoring the health situation of field is very strong,it is worth to do further research.In the distributed optical fiber sensing system,when the sensing optical fiber stimulated by stress or temperature change,part of the system parameters will also change,to change the parameters and the matching is obtained by signal processing to the change of stress or temperature.In view of distributed optical fiber sensing system,this paper makes the following research:firstly,fully understand the background and significance of the project,and analyze the demand analysis;Possess the theoretical basis of distributed optical fiber sensing system based on brillouin scattering;Then,the optimization algorithm of brillouin scattering spectrum feature extraction is proposed and the numerical analysis is carried out.Finally,the experimental system is built to verify the effectiveness of the optimized algorithm by experimental data.Using compression factor of particle swarm algorithm?YSPSO?reverse adjustment of the weights of neural network?BP?method was carried out on the brillouin scattering spectrum feature extraction,effectively overcome the traditional BP neural network iteration slow speed and low accuracy of shortcomings,improve the accuracy of the amount of the extraction of brillouin frequency,guarantees the solving speed.Genetic algorithm?GA?was proposed for the combination of quantum particle swarm optimization?QPSO?hybrid optimization algorithm,the optimization algorithm has such features:GA algorithm better control the QPSO algorithm the balance between global search and local search,increase the information interaction between the particles,and effectively improve the extraction precision of the brillouin scattering spectrum.The results of numerical analysis shows that the simulation fitting degree?R²?of different lines is above 0.98,and the fitting mean square error is in the order of 10^-5orders of magnitude,and the two optimization algorithms fit well.Build the BOTDA systemin the lab,the optical fiber to be included in the water bath pot heating to realize the rise of temperature,using the vector network analyzer measured brillouin scattering spectrum,resulting in different temperature of brillouin scattering spectrum data,using the proposed two algorithms for processing experimental data,the results shows that both algorithms can improve the precision of brillouin scattering spectrum feature extraction,under 25?,YSPSO-BP algorithm fitting error is 2.05 MHz,GA-QPSO algorithm fitting error is 2.13 MHz.when the temperature rise up,the fitting error is reduced.At 85?,YSPSO-BP frequency shift fitting error is 0.045 MHz,GA-QPSO fitting method of error of 0.051.Therefore,two new algorithms are applied to brillouin scattering temperature strain sensing system,which can improve the detection accuracy.