Temperature Measurement Of Brillouin Optical Time Domain Analysis System Based On Support Vector Machine

The Brillouin optical time domain analysis(BOTDA)system has been extensively studied in recent years for its advantages in distributed sensing.Brillouin optical time domain analysis system is widely used in the health monitoring of some large facilities for it does not require complex processing and packaging for optical fibers,transmission and sensing integration and low testing cost.In order to improve the test accuracy of the system,the support vector machine(SVM)is used to extract the temperature information in the Brillouin optical time domain analysis system.In the training phase,the ideal Voigt curve is used as the Brillouin gain spectrum to train the temperature values of each category at the same time,and the support vector machine model is obtained,and the Brillouin gain spectrum under different conditions is studied.The results show that compared with the traditional algorithm,the support vector machine model can improve the accuracy of temperature extraction and reduce the running time of the algorithm in the larger signal-to-noise ratio,pump pulse width and sweep step size.The measurement accuracy is improved while the data processing time is shortened.Based on support vector machine temperature extraction of Brillouin optical time domain analysis system,three algorithms(particle swarm optimization,genetic algorithm and firefly algorithm)are used to optimize the support vector machine,and the temperature information extraction in Brillouin optical time domain analysis system is realized.In the training phase,the Voigt curve is still used as the Brillouin gain spectrum for each category of temperature values,and the optimized support vector machine model is used for data classification.By studying the Brillouin gain spectrum under different conditions,the results show that the optimized SVM obtains the higher temperature extraction accuracy in the range of larger signal-to-noise ratio,pump pulse width and sweep step.In addition to higher precision,there is no significant loss in processing time,and the data processing speed is greatly improved compared with the traditional Lorentz algorithm.The fast processing speed and better accuracy make the optimized SVM have better application significance in the application of BOTDA sensor.