Improving the peak wavelength detection accuracy of Sn-doped, H 2-loaded FBG high temperature sensors by wavelet filter and Gaussian curve fitting

Abstract

Fiber Bragg grating (FBG) written into Sn-doped, H 2-loaded fiber could be used as high-temperature sensors up to 800 °C, but the peak wavelength detecting accuracy was unsatisfactory due to the low signal-noise ratio (SNR) and spectrum distortion. In this paper wavelet filter and Gaussian curve fitting were jointly applied to improve the wavelength detection accuracy of this kind of novel sensor. Firstly wavelet filter was used to suppress the noise in the noise-contaminated FBG reflection spectrum. The denoising efficiency of different wavelet functions and decomposition levels were evaluated by calculating the SNR of denoised signals, and dmey wavelet with decomposition level of 9 presented the best result. Then the denoised signals were fitted to Gaussian profile and the goodness of fit was compared with those obtained without wavelet denoising. Simulation and experimental results demonstrated that the joint application of wavelet filter and Gaussian curve fitting was a promising approach to enhance the peak wavelength detection accuracy of FBG high temperature sensors.