Research on temperature compensation method for photoelectric sensors

Abstract

In order to make the temperature error of fiber-optic current sensor (FGB) based on polarization modulation principle meet the requirements of engineering applications, the temperature error characteristics of FGB are analyzed theoretically, and the optimized BP neural network is used for the temperature compensation of FGB, which realizes the nonlinear temperature error correction of the sensor, and compares and analyzes the experimental results with those of other types of temperature compensation algorithms. The results show that the temperature compensation results based on the neural network algorithm are better than other compensation effects. Finally, the repeatability of the FGB was experimentally verified using its full temperature experiment, and the temperature errors of the FGB in the range of 20 ℃~ 100 ℃ were less than 0.5% after the correction of the neural network algorithm.