A Method of HCN Gas Spectrum Denoising and Baseline Removal Used for FBG Interrogation

Abstract

Hydrogen cyanide (HCN) gas based fiber Bragg grating (FBG) interrogation is a promising approach in thermal cycling environment. However, the HCN spectrum suffers from strong noise and baseline wander induced by the interrogation system, which limits the measurement stability and reliability. In this paper, we propose an integrated method for spectral denoising and baseline removal to improve the interrogation performance. The method combines the classical empirical mode decomposition (EMD) with the recently developed convolution window (CW) filtering. After the HCN spectrum is decomposed by EMD, the noise and baseline components are extracted from the intrinsic mode functions via a “multiband” filtering approach. Verification experiments are performed using HCN gas cells with different absorption spectra. The results show that the proposed method is able to remove both noise and baseline with minimum signal distortion. Compared with other methods, the optimal interrogation stability with 25 Torr HCN gas cell is improved from ±1.43 pm to ±0.87 pm. For the 100 Torr HCN gas cell, the optimal interrogation stability is improved from ±1.75 pm to ±0.99 pm. In addition, the proposed method also improves the reliability of the interrogation system in the case of large light attenuation.