- Biomedical Optics & Medical Imaging
- Defense & Security
- Electronic Imaging & Signal Processing
- Illumination & Displays
- Lasers & Sources
- Micro/Nano Lithography
- Optical Design & Engineering
- Optoelectronics & Communications
- Remote Sensing
- Sensing & Measurement
- Solar & Alternative Energy
- Sign up for Newsroom E-Alerts
- Information for:
Sensing & Measurement
Pushing the limit on the distributed Brillouin sensors
A technique that exploits different pulse widths delivers the best spatial resolution and sensing length to date for distributed sensors based on scattering of light by sound waves.
16 July 2010, SPIE Newsroom. DOI: 10.1117/2.1201006.002939
Twenty years have passed since Brillouin optical time domain analysis (BOTDA) was first used to detect 3°C temperature resolution and 100m spatial resolution over a 1.2km optical fiber (the Brillouin frequency is proportional to temperature and strain).1–3 These ‘distributed’ sensors have enormous potential for monitoring the health of large civil structures such as dams, bridges, and pipelines, which require long sensing length for coverage in two and three dimensions. Because of their ability to locate and identify microcracks and other anomalies, distributed sensors can help to prevent disasters due to corrosion of steel and concrete interfaces, deformation of pipelines, and gas, oil, and water leaks in pipes. However, achieving high-performance distributed sensors with high spatial resolution (<1m) and high strain resolution (10με) over long sensing lengths (>10km) poses a considerable challenge.
Early development of distributed Brillouin sensors over distances4,5 employed a loss rather than a gain technique.3 In the gain process, the pumped pulse transmits energy to the probe wave through stimulated Brillouin scattering, and the pump wave can quickly be depleted. Brillouin loss, however, uses a continuous wave as pump, which is more robust. In 1995, we achieved the longest sensing length to date of 51km, with 5m spatial resolution and 1°C temperature resolution.5 Figure 1(a) shows the Brillouin gain signal and Figure 1(b) the linewidth of the Brillouin gain versus the spatial resolution from these experiments.6
(a) Peak Brillouin loss as a function of pulse duration.7
(b) Brillouin spectral width vs. pulse width.8
It is very clear that shorter spatial resolution results in low gain: see Figure 1(a). Indeed, achieving a high signal-to-noise ratio requires the narrow spectral width typical of a large pulse: see Figure 1(b). But high power and greater pulse width distort the Brillouin spectrum. Consequently, lower power is preferable.
The spatial resolution of the distributed Brillouin scattering fiber sensor is limited to 1m, which is equivalent to the relaxation time of an acoustic wave (10ns). When a short pulse (<5ns) is used for submeter spatial resolution, the Brillouin gain becomes very weak. The broadened Brillouin spectrum determined by the convolution of the pulse and natural Brillouin gain spectrum prohibits short pulses, except for those with a small DC level that provides prepumping for the acoustic field.7,9 This produces a Brillouin spectrum whose width is close to the natural Brillouin spectrum. However, the DC-level-induced prepumped acoustic wave tends to deplete the Stokes signal in the Brillouin loss-based sensor over the long sensing length, which limits the use of prepumping for such applications.
Because of the high gain in the long pulse, we used two large pulses with slightly different pulse widths to measure BOTDA in the time domain. Then we subtracted the results of two different pulses at each scanned Brillouin frequency to give the differential gain at every fiber position. This new technique, called differential pulse-width pair (DPP)-based BOTDA,10,11 is shown schematically in Figure 2, where δτ is the pulse-width difference, and δz is its equivalent length. This approach has been used to demonstrate 25km sensing length on dispersion-shifted fiber and submeter (50cm) spatial resolution11,12 for a strain resolution of 10με (i.e., 10μm strain over a length of 1m).
Figure 2. Working principle of differential pulse-width pair Brillouin optical time domain analysis (DPP-BOTDA) sensor. I: Intensity. 0: Initial fiber point. t: Time. τ1,2: Pulse width. δτ: Pulse-width difference. δz: δτequivalent length. c: Speed of light in fiber.
Achieving even longer sensing lengths (>30km) calls for coded pulses to meet the requirement of low power. Using DPP-BODTA with an RZ (return-to-zero)-coded pulse and a 55/60ns pulse pair, we measured a stress length of 50cm over 50km sensing length (see Figure 3). The uncertainty of the Brillouin frequency shift is 0.7MHz,13 which is proportional to a strain resolution of 12μm. These findings represent the best results obtained thus far (and the best balance of trade-offs) for the combination of sensing length with spatial and temperature resolution using Brillouin scattering-based distributed sensors. We plan to further develop extended sensing length by adding inline fiber amplifiers, which help to sustain a high signal-to-noise ratio over great distances.
Figure 3. The differential Brillouin spectrum of RZ-coded pulses at the 50cm stress position over a sensing-fiber length of 50km. BFS: Brillouin frequency shift.
University of Ottawa