Research On Wavelength Scanning Nonlinear Suppression Of Tunable Filter For FBG Demodulation System

Due to small size,light weight,high sensitivity,easily multiplexing,and strong anti-electromagnetic interference capability,the fiber Bragg grating(FBG)sensors have been widely used in various fields in recent years.The fiber grating sensing demodulation method based on F-P tunable filter is the key technology in the field of fiber sensing demodulation.But the F-P filter performs nonlinear scanning wavelength due to piezoelectric ceramic characteristics,which affects the stability of demodulation results.In this paper,based on the FBG demodulation scheme of F-P tunable filter,a fiber grating sensing demodulation system based on long optical path difference Michelson fiber interferometer is proposed.The fiber grating demodulation algorithm based on BP neural network is also proposed and verified.The specific work completed is as follows:1.The fiber grating sensing demodulation system based on long optical path difference Michelson interferometer and the demodulation algorithm under nonlinear scanning are proposed.The scanning light source module is implemented.The Michelson interferometer is designed.Compared with the traditional broadband light source,the scanning light source has the characteristics of narrow line width and long coherence length.The Michelson interferometer has an optical path difference of 73.4mm,and between the adjacent transmission peaks of the F-P etalon with FSR of 100 Ghz,up to 48 interference fringe extreme points can be inserted for the wavelength reference.2.The experimental results of stable and variable temperature environments verify the suppression effect of long optical path difference Michelson interferometer on the scanning nonlinearity of tunable filters.Experiments show that at an average temperature change rate of 6 °C/min,the etalon-based demodulation fluctuation amplitude is ±20.39 pm,the standard deviation is 7.08 pm,and the interferometer-based demodulation wavelength fluctuation amplitude is ±6.89 pm,standard deviation is 3.11 pm.The experimental results verify that the interferometer can effectively suppress the filter nonlinearity in the variable temperature environment and improve the stability of the sensor demodulation system.3.The basic structure and algorithm principle of neural network are studied.The fiber grating sensing demodulation algorithm based on BP neural network is proposed and applied to FBG sensor demodulation system based on etalon.The four-layer BP neural network for demodulating the FBG Bragg wavelength is constructed,which takes the scanning wavelength sampling point of the F-P filter as the input,and the reference wavelength as the output.Experiments show that the proposed algorithm has a demodulation accuracy of ±1.65 pm at stable temperature and a standard deviation of 0.62 pm,which is superior to linear interpolation and cubic spline interpolation.It proves that BP neural network can effectively improve the demodulation accuracy in FBG demodulation.Through the above work,from both the hardware system and the demodulation algorithm,in constant temperature and variable temperature environments,the demodulation error caused by the non-linearity of the filter scanning is effectively suppressed.