The Research On The Method Of Wavelength Detection Based On Fiber Bragg Grating Sensor

Because of the many advantages, optical fiber sensing technology has been widelyused in the areas of military, aviation, healthcare and chemistry. Demodulation is one ofthe key technologies in fiber Bragg grating (FBG) sensor system. Meanwhile, theemphasis of the demodulation is the wavelength detection technology, including withde-noising and peak searching. Select proper de-noising scheme and peak detectingstrategy is the basic guarantee which can improve demodulation accuracy. This aspectof the study has certain academic significance and engineering application value.This paper adopt the wavelet threshold method to de-noise for noisy signal. Twokinds of de-noising schemes of FBG signal based on the wavelet threshold method aredesigned after analyzing the four factors that affect wavelet de-noising, which includesdecomposition scale, wavelet basis, threshold and threshold function.①One is dealingwith the wavelet coefficients term by term. Proposing the optimal decomposition scalemethod based on wavelet energy entropy and weighted fusion wavelet base method. Asfor wavelet de-noising threshold and threshold function, this article not only improvesthe self-optimizing threshold selection strategy, but also put forward the method ofselecting the optimal value in compromising threshold function.②The other is topresent a revised threshold and threshold function which can change along with theincrease of decomposition scale in terms of neighborhood coefficient.In order to obtain accurate center wavelength drift, a FBG peak detection schemenamed three-point peak-detection based on sparse spectrum is proposed by analyzingthe advantage and disadvantage of existing peak-detection algorithm. On that basis, thethree factors which influence the accuracy of peak detection are determined, includingwith the window size, the sparse spectrum data and wavelength interval. In this paper,method of derivation is used to determine the spectrum processing window size.Simultaneously the article set the sampling interval to resample the window spectrum,so as to obtain sparse spectral data and realize polynomial fitting.The fitting results ifuseful for us to select suitable wavelength interval, then we could determine threecoordinates and finish the peak detection of FBG reflection spectrum.The platform is established based on FBG temperature-sensing system. Theeffectiveness of de-noising method and peak-detection algorithm proposed in this paper is verified by experiments.