A fast key parameter extraction algorithm for long fiber distributed sensing based on Brillouin scattering.

Abstract

A fast key parameter extraction algorithm is proposed to improve the real-time performance of temperature and strain measurements when performing Brillouin scattering-based fiber-distributed sensing. The algorithm uses a new initial value method that takes the extracted key parameters of the current point in the fiber as the initial guesses for the next point. Based on the old and new initial value method, the existing objective method, optimization algorithm, and convergence criterion, the key parameter extraction algorithms developed are implemented in Matlab using the typical Lorentzian, Gaussian, and pseudo-Voigt profiles. These algorithms are used to extract the parameters over a large range of measured Brillouin spectra for the entire fiber with different averaging times. The results reveal that apart from the case when the frequency sweep spans is less than the linewidth and the pseudo-Voigt profile is used (in this case, the mean computation time of the proposed algorithm is 1.1% larger than that of the referenced algorithm), the proposed algorithm not only ensures high accuracy in extracting the key parameters, but also improves the arithmetic efficiency by 16.3%-49.1%.