Temperature Sensing Research Based On DBR Fiber Laser Beat Detection Combined With BP Neural Network

Fiber Bragg grating sensing technology has been widely used in military,electricity,petroleum,petrochemical,transportation,construction and other industrial fields.Its sensing and demodulation method has always been the focus of attention.The common commercial method is to use optical fiber F-P cavity scanning optical interference demodulation,these methods are complex and costly.In order to reduce the system cost,a method of heterodyne demodulation is proposed.This technique is widely used in distributed feedback Bragg fiber laser(DFB)structures,using its birefringence in the detector to form the beat detection demodulation.In this method,it is more complicated to achieve stable beat frequency.Then,it is proposed to use the distributed reflection laser(DBR)structure,the use of fiber laser cavity length changes to form a multi-longitudinal mode beat frequency sensor demodulation.In this paper,a fiber grating temperature sensing and demodulation method based on laser beat measurement is proposed,and a three-layer BP neural network model is constructed to optimize the temperature sensing data.The method uses linear chirped grating(CFBG)sensor and fiber grating(FBG)as the feedback mirror fiber laser system,and uses the grating temperature sensor temperature sensor to measure the laser beat frequency.In the past,we often use chirped fiber grating(CFBG)delay characteristics as a reference standard,assuming that the chirped fiber grating(CFBG)linear time delay is ideal,but in practical applications,due to production process constraints,chirped fiber grating CFBG)is not completely linear time delay while jitter.According to the linear delay test results,there are obvious systematic errors.Therefore,in order to reduce the system error,the beat frequency should be calibrated according to the specific time delay characteristic curve of the chirped fiber grating(CFBG).Because the temperature measurement error reduces the CFBG nonlinear delay and delay jittercaused by the three-layer BP neural network model,the BP neural network algorithm has good fault tolerance and nonlinear mapping ability,and can approximate any nonlinear function to solve the complex nonlinear relationship parameter.In the experiment,the measurement was repeated 10 times,and 10 sets of beat frequency and temperature data were obtained.Randomly select the 9 groups as the training frequency data set,the input layer and the three-tier BP network model as the network input value,the actual temperature as the corresponding value of the network output value,training the network parameter until the parameters and network structure optimization.The other group is used to validate the sample set to test the usefulness of the network model,the temperature sensitivity and correlation coefficients of this set of data were 37.89 KHz /? and 99.767%,respectively,and their correlation coefficient was 99.95%.The results show that the measurement accuracy of the system can be greatly improved by using the three-layer BP neural network algorithm.In this paper,based on BP neural network algorithm to achieve temperature sensing,the beat frequency demodulation technology was more practical.