Benefits of Spectral Property Engineering in Distributed Brillouin Fiber Sensing

Abstract

As one of the most consolidated distributed fiber sensors based on stimulated Brillouin scattering, the Brillouin optical time-domain analyzer (BOTDA) has been developed for decades. Despite the commercial availability and outstanding progresses which has been achieved, the intrinsic Lorentzian gain spectrum restricts the sensing performance from possible further enhancements and hence limits the field of validity of the technique. In this paper, the novel method of engineering the gain spectral properties of the Brillouin scattering and its application on static and dynamic BOTDA sensors will be reviewed. Such a spectral property engineering has not only provided improvements to BOTDA, but also might open a new way to enhance the performance of all kinds of distributed Brillouin fiber sensors.

Keywords: Brillouin optical time-domain analyzer; distributed fiber sensing; slope-assisted Brillouin optical time-domain analyzer; spectral property engineering; stimulated Brillouin scattering.

Figure 1

Engineered BGS by the superposition of a gain with two symmetric losses.

Figure 2

(a) Typical conventional (red) and engineered (blue) simulated BGS ($d=1$) with the same noise level and their corresponding fitting curves. The conventional BGS is excited by a 100 ns pump pulse with a spectrum FWHM of 54 MHz; (b) the magnified peak area of the conventional BGS (rectangular area in (a)), highlighting the BFS estimation error $\Delta f_{ci}$ due to the noise in the ith measurement.

Figure 3

Simulation results of: (a) the frequency error ratio $\xi$ as a function of m and d; (b) the distribution of the BFS determination error with the conventional and engineered BGS with selected m and d values in (a) marked as $P1$–$P3$.

Figure 4

Schematic of (a) multi-pump wave and (b) multi-probe wave scheme. Red arrows mark the pump and gray arrows the probe waves.

Figure 5

Experimental results of: (a) the reconstructed conventional (red) and engineered BGS with (gray) and without (blue) gain recovery; (b) the estimated BFS distribution with a stretching section.

Figure 6

The frequency error distribution measured by the conventional (red) and engineered BGS with (green) and without (blue) gain recovery.

Figure 7

Simulation results of: (a) the conventional (red) and engineered (blue) BGS with the schematic principle of SA-BOTDA at an arbitrary interrogation point (green dot); (b) the slope of the conventional (red dashed) and engineered (blue dashed) BGS in the shadowed frequency area in (a).

Figure 8

(a) Schematic spectrum of the multi-probe wave scheme and (b) specific structure of the sensing fiber.

Figure 9

Experimental (a) conventional (red) and engineered (blue) BGS and (b) their slopes as a function of the interrogation frequency.

Figure 9

Experimental (a) conventional (red) and engineered (blue) BGS and (b) their slopes as a function of the interrogation frequency.

Figure 10

Sinusoidal Brillouin gain response of the conventional (a) and engineered (b) BGS at the interrogation frequencies in Figure 9a.

Figure 11

FFT spectra of the conventional (a) and engineered (b) BGS at the interrogation frequencies in Figure 9a.

Figure 12

Relative second order harmonic level as a function of the interrogation frequency. A-F are the interrogation frequencies in Figure 9a.

Figure 13

Time domain traces measured with the conventional (red) and engineered (blue) BGS at a pump-probe frequency offset of 10.605 GHz for static sensing. The voltage level of the trace with the engineered BGS is down shifted for a better visualization.

Figure 14

Noise level of the detected trace baseline as a function of the interrogation frequency for dynamic sensing.