An Improved Temperature Compensation Method for Fiber Bragg Grating Pressure Sensor Based on Extreme Learning Machine

Abstract

According to the problem of the sensor nonlinear changes occur at high temperatures, extreme learning machine model, is presented in this thesis the pressure sensitive grating and removing the temperature of the grating experiment data for training, establish a nonlinear model of wavelength, temperature, predict the experimental temperature, then the temperature data of pressure-sensitive grating the training set of training samples, the nonlinear model, temperature - wavelength prediction test set sample output wavelength, achieve the goal of improved temperature compensation method. The experimental results show that the algorithm can achieve a more ideal temperature compensation effect.