An EMD-SG Algorithm for Spectral Noise Reduction of FBG-FP Static Strain Sensor

Abstract

Random noises in the reflection spectra of fiber Bragg grating (FBG) typically limit the resolution of static-strain sensor based on FBG. We use the empirical mode decomposition (EMD) and the Savitzky–Golay (SG) filter to denoise reflection spectra from a pair of FBG-based Fabry–Perot interferometers (FBG-FPs) for achieving high static-strain resolution. The EMD is proposed for extracting adaptive basis functions from two FBG-FPs’ reflection spectra. The SG filter is proposed for smoothing and denoising basis functions. These basis functions are used to reconstruct the reflection spectra, which can be used to calculate the relative wavelength difference. A simulation and two experiments are conducted to verify the performance of the proposed technique. The results of experiment conducted in the noisy laboratory and in the quiet basement are consistent with the simulation result. When SNR is less than 10 dB, the resolution of the EMD-SG method is approximately ten times the resolution of the traditional cross correlation. A minimumstatic strain resolution of 0.68 $\text{n}\varepsilon $ is achieved.