We propose and experimentally demonstrate a scheme of polarization independent fast Brillouin optical time domain analysis (F-BOTDA) based on pump frequency modulation and cyclic coding. The Brillouin gain spectrum (BGS) is reconstructed by fast scanning frequency of the pump using an arbitrary waveform generator (AWG). To realize long range distributed dynamic strain sensing, polarization diversity technique and cyclic coding are employed to eliminate polarization fading and enhance the signal-to-noise ratio (SNR). Based on this configuration, the need of trace averaging is avoided, sensing speed of 440 Hz is achieved over ~2 km single mode fiber with 50 scanning frequencies and a spatial resolution of 1.5 m. Vibration events up to 40 Hz are successfully identified.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Distributed optical fiber sensor (DOFS) based on Brillouin scattering has been extensively studied and discussed for structural health monitoring in diverse fields over the past two decades, due to its advantages of dielectric, small in size, immune to electromagnetic interference and high sensitivity [1–5]. For Brillouin scattering based DOFS, there are mainly three schemes to realize distributed sensing: Brillouin optical time domain analysis/reflectometry (BOTDA/R), Brillouin optical correlation domain analysis/ reflectometry (BOCDA/R) and Brillouin optical frequency domain analysis (BOFDA). Among them, BOTDA/R have distinct superiority in sensing range. With the help of advanced coding and distributed amplification technique, measurement range up to 100 km has been demonstrated with a meter-scale spatial resolution [6–9]. Compared with BOTDA/R, BOCDA/R possess high spatial resolution, but a limited sensing range [10-11]. On the other hand, the sensing speed of Brillouin scattering based DOFS is fairly slow due to time-consuming frequency sweeping and larger number of averages to improve the signal-to-noise ratio (SNR), which makes it more suitable to static parameters measurement rather than dynamic detection .
In order to solve this problem, various techniques on dynamic Brillouin sensing have been presented in recent years. For instance, coherent detection and advanced image denoising have been introduced into BOTDA to substantially improve the SNR and avoid trace averaging [13, 14]. Tur et al. have proposed slope-assisted BOTDA (SA-BOTDA), in which the frequency of probe light is fixed at the half of BGS to avoid frequency sweeping . The vibration-induced modulation of the local BFS is translated into an intensity variation of the probe light along the sensing fiber. Voskoboinik et al. have developed sweep free BOTDA (SF-BOTDA) [16–18], which employs multiple probes to interact with multiple pumps simultaneously to reconstruct BGS, where each of pump-probe pair replaced one sweeping step in the classical BOTDA technique. Much higher sensing speed have been demonstrated by single-shot BOTDA in which digital optical frequency technique was involved, all the temperature or strain information can be obtained through only one-shot measurement . More recently, a novel optical chirp chain technique was proposed for a high-speed BOTDA realizing an extremely fast dynamic strain measurement with a MHz-order sampling repetition . Using BOCDR to realize fast sensing was proposed by Mizuno , where a strain sampling rate of up to 100 kHz at an arbitrary position was experimentally verified. While for distributed measurement, the repetition rate was reduced to 100 Hz due to the need of position shift. Alternately, fast BOTDA (F-BOTDA) still involves frequency sweeping to reconstruct BGS, however, employing high-performance arbitrary waveform generator (AWG) to reduce frequency switching time of probe. Vibration sensing of a 100m fiber with a sampling rate of ~8.3 kHz was demonstrated by using a combination of a dual-channel AWG together with a microwave vector signal generator with I/Q inputs . Then spatial resolution of F-BOTDA was improved to 20 cm by using differential pulse-width pair and an AWG with ~5.5 GHz bandwidth . By employing the polarization-independent technique, full scans of the BGS on a 145-m long fiber at an acquisition rate of 9.7 kHz was also reported . F-BOTDA system provides an impressive performance in both acquiring speed and spatial resolution, while its sensing range is commonly limited to about several hundred meters. Longer distance means poorer SNR, the number of averaging will become the dominant factor to deteriorate the sensing speed of fast BOTDA.
In this paper, we demonstrate a polarization independent fast BOTDA based on pump frequency modulation and cyclic coding. Both the pump and probe performed double sideband (DSB) modulation, while the frequency of probe keeps constant, the frequency of pump is swept by using AWG to fast scan the Brillouin spectral response. To further reduce the number of average, polarization diversity technique and cyclic coding are also employed. Benefiting from the advanced techniques above, the need of trace averaging is avoided and the sensing range is extended to the order of kilometer. Dynamic strain measurements are implemented in a 2 km single mode fiber. Vibration events up to 40 Hz are demonstrated with a sensing speed of 440 Hz and 1.5 m spatial resolution.
The conceptional schematic of F-BOTDA based on pump frequency modulation is shown in Fig. 1. Compared with the proposed fast BOTDA previously, the frequency of pump is modulated to scan the Brillouin spectral response, while the optical frequency of the probe is kept constant, hence, fast frequency switching could be achieved by using AWG with a few hundred MHz bandwidth. The modulation waveform of pump can be expressed as:
Where V0 is the amplitude of AWG output, τp is the width of pump pulse, Tin is the interval between two pump pulses which is determined by round-trip time of flight through the fiber under test (FUT). fi (i = 1...M) is the frequency of the ith frame, f0 is the onset frequency and fstep is the scanning step of frequency. In this way, the pump wave is frequency modulated and pulsed by only using one EOM. Note that f0 should be large enough to avoid both the two tones of pump interacting with probe wave simultaneously. The upper sideband of probe will interact with the upper sideband of pump via loss process, while the lower sideband of probe will interact with the lower sideband of pump via gain process as shown in Fig. 1(a). To avoid polarization fading due to the strong polarization sensitivity of stimulated Brillouin scattering (SBS) and reduce the number of averages, the polarization states of two sidebands of probe are set to be orthogonal. Assuming that the Brillouin gain is small, after interacting with pump pulse, the power of each sideband of probe at a location z follows:
Where Iup(L,Δv) and Ilow(L,Δv) are the intensity of upper tone and lower tone of probe at the far end of the fiber (z = L), Δv is the local detuning of the probe wave frequency from the center of the Brillouin spectrum, a is the fiber attenuation, ηup and ηlow are the mixing efficiency factor due to polarization state mismatch of the pump and probe. Δz is the spatial resolution (proportional to the pulse width τp), gB(z,Δv) is the Brillouin gain, Ip(L,Δv) is the intensity of pump at a location z. For an amplitude-modulated scheme, Iup(L,Δv) = Ilow(L,Δv). After balanced detection, the overall signal can be described as:
Since the polarization states of Iup(L,Δv) and Ilow(L,Δv) are orthogonal, ηup + ηlow should equal to 1 in theory. Therefore, the polarization fading in time trace can be canceled. Also, the larger useless DC component will be removed, as well as the quantization error in analog-to-digital conversion process.
In order to further reduce the number of averages and improve the sensing speed, we employ cyclic coding in F-BOTDA system due to its unique real-time decoding property. For a code length of L, the code is generated following the recurrence equation for an integer n :
3. Experimental setup
The experimental setup for polarization independent fast BOTDA based on pump frequency modulation and cyclic coding is shown in Fig. 2. A narrow linewidth distributed feedback (DFB) laser (NKT Laser, E15) operating at 1550.13 nm is used as light source. The output of the laser is divided into two propagation paths by using a 3dB optical coupler. The upper branch signal is used as the probe and the lower branch signal is used as the pump. The polarization controller (PC) is employed at the input of each MZM to achieve the maximum modulation efficiency. On the lower branch, microwave pulse trains generated from the AWG (Tektronix, AWG5012C) are modulated onto the pump light through the high extinction ratio electro-optic modulator (EOM, >40dB) biased at null-transmission point. The pump pulses are amplified by erbium-doped fiber amplifier (EDFA) before launched into the FUT.
On the upper branch, an EOM and a radio-frequency generator are used to perform double-side-band (DSB) modulation. Then EDFA is used to amplify the power probe. The upper sideband and lower sideband are separated by using a fiber Bragg grating (FBG) with 0.2 nm bandwidth. Their polarization state is adjusted by two PC, thereby enabling the same power after recombining through a polarization beam combiner (PBC). Although there would be a delay between the two sidebands, this delay is negligible since the difference of the transmission path is only several centimeters.
After interaction with the pump signal, the probe wave is divided by another narrow bandwidth FBG and detected by a balanced photon detector (PD) with 350-MHz bandwidth. The electrical signals output from the PD are collected by a digital data-acquisition (DAQ, Gage, 4GSa/s maximum sampling rate) card with 1 GHz sampling rate (corresponding to 0.1 m sampling interval) and the collected data are then sent to a computer for further processing.
1.975 km single mode fiber whose BFS is 10.721 GHz at room temperature is used as FUT. The layout of FUT is shown in Fig. 2(b). Near the end of the fiber, a section of 2 m is stretched by two translation stages.
4. Experimental results and discussion
4.1 Static measurement
Firstly, the spectra of probe light were measured by using an optical spectrum analyzer (OSA, AQ6370C) as shown in Fig. 3. The blue line and red line show the upper tone and lower tone of probe wave (according to frequency) before PBC. It can be observed that they are separated by 0.178 nm with a sideband suppression larger than 10.5 dB. The small peak near carrier wave is caused by the harmonic noise of microwave generator. The yellow line shows the recombined probe light after PBC. The power of the two tones are set to be equal by adjusting PC on each path.
In the static measurement, the duration of the pump pulses τp is set as 15 ns, corresponding to 1.5 m spatial resolution. The peak power of pump pulses is controlled to 20 dBm, to prevent nonlinear effects such as modulation instability. While the power of each probe sideband is kept at −10 dBm in order to avoid the non-local effect. Tin is set as 22.72 µs, larger than round-trip time of FUT, f0 is 100 MHz to avoid BGS distortion, fstep is 3 MHz, frequency number is 50, enabling a dynamic range of 147 MHz. Figure 4(a) and Fig. 4(b) plot the Brillouin time-domain traces with single sideband detection at frequency of 10.721 GHz with an average number of 200. It can be observed that when the lower tone of probe experiences the maximum response, the upper one experiences the minimum response. Therefore, the polarization fading in the balanced detection is mostly removed.
Figure 5(a) shows measured BGS distribution along the whole FUT with an average number of 50. The leading fiber and stretched fiber near the end can be clearly observed. With the help of Lorentzian curve fitting, the distribution of BFS around the last 30 m fiber is illustrated in Fig. 5(b). Figure 5(b) inset gives the BFS distribution along the FUT. The BFS uncertainty of the system, defined as the standard deviation of the BFS values around the last 30 m fiber, is calculated to be 0.58 MHz. And the BFS difference between the stretched and unstretched fiber is found to be 78 MHz. The ~1.5 m transition section also implies a spatial resolution of 1.5 m, agrees with 15 ns pump pulse well.
Distributed strain sensing experiment is also carried out to test our F-BOTDA performance. One of the translation stages is shifted by 0.5mm corresponding to a strain step of 250 με, the experiment results shown in Fig. 6. When strain grows larger, the BFS shifts upwards. Then we record the BFS of the stretched position and evaluate its relationship with the strain via linear fitting, as shown in Fig. 6(b). The strain coefficient for the FUT is 0.0432 MHz/με. The coefficient of determination R2 is 99.69%, indicating that good linear relationship is achieved for strain sensing.
In order to further reduce the number of averages and improve the sensing speed of our F-BOTDA system, 71-bit cyclic coding is employed [26–28]. To enable a return-to-zero (RZ) pump modulation format, the pulse width is still 15 ns and bit duration is set to 320 ns. Figure 7(a) shows the measured BGS distribution along the FUT for 71-bit cyclic coding without averages. Figure 7(b) illustrates the obtained time traces at the frequency shift 10.721 GHz when cyclic coding is used or not. It can be clearly seen from the figure that both traces are characterized by approximately the same SNR, meaning that the number of averages can be significantly reduced by using cyclic coding. It should be noted that although the cyclic coding trace is without averages, however, the coding pulse need to spread the whole FUT, therefore, an extra round-trip time is required. The acquisition time of each frequency will be twice of round-trip time. Thus, for a scanning frequency number of 50, the BGS acquiring time should be equal to 50 × 2 × 22.72 us, corresponding to a sampling rate of 440 Hz.
4.2 Enhance dynamic response by using cyclic coding
In vibration measurement, the stretched fiber was vibrated periodically by an electric-motor-driven eccentric wheel to imitate vibration signal. Figure 8(a) is the measured BGS as a function of time at loose fiber for reference. The results are moving averaged of order 5 to improve SNR. Figure 8(b) shows the evolution of BGS at the vibrated section when the driving voltage of the motor is 1.9 V. From Fig. 8(b) the vibration signal can be clearly observed, the period and amplitude are 32 ms and 195 με, respectively. The higher frequency vibration signal was demonstrated by raising the driving voltage to 2.4 V, as shown in Fig. 8 (c). It also observed a non-sine signal of the strain variation, which should be caused by the non-uniform vibration of the eccentric wheel. Signal up to 220 Hz is expected to be obtained since the effective sampling rate of the system is 440 Hz. The BFS uncertainty is measured to be 0.74 MHz, corresponding to a strain accuracy of 17 με.
To determine the accurate vibration frequency, the vibrate traces in Fig. 8 are analyzed by using the Fast Fourier Transform (FFT), the frequency domain spectrum are shown in Fig. 9. The fundamental frequency of 31 Hz, 41 Hz and the second-order harmonic can be clearly identified.
5. Discussion and conclusion
For BOTDA, the sensing speed is mainly limited by four factors: 1). the round-trip time of pump pulse flight through FUT, 2). the granularity of the frequency sampling of the BGS, 3). the switching speed of the optical frequency, 4). the number of trace averages. Factors 1 and 2 impose an ultimate bound on the acquiring speed of F-BOTDA, which cannot be improved. For the proposed scheme, the frequency switching time is reduced to the order of ns with the help of AWG. Meanwhile, the need of trace averaging is avoided by using polarization diversity and cyclic coding technique. Compared with previously presented F-BOTDA, our proposed scheme has extended the sensing range to ~2 km with a considerable sampling rate. The overall measurement time includes the acquisition time and the data processing time. The acquisition time is about 8 s in our dynamic experiment, mainly contributed to the data fetching and saving time (The data collection program is based on Matlab). In conclusion, we have proposed a polarization independent F-BOTDA system based on pump pulse frequency modulation and cyclic coding. The frequency of pump is swept by using a low bandwidth AWG to cover the BGS. Meanwhile, the two orthogonal polarized sidebands of the probe are detected by balanced detector to eliminate polarization fading. Combining with cyclic coding technique, the sensing range of F-BOTDA has extended to the order of km without the need of averaging. For a 2 km single mode fiber and 50 scanning frequencies, dynamic measurement was experimentally demonstrated with a maximum sensing speed of 5.8 × 105 BGS/s. The vibration signals up to 40 Hz have been observed, where the spatial resolution and strain resolution are 1.5 m and 17 µε, respectively.
Key Research and Development Project of Ministry of Science and Technology (No. 2016YFC0801200); Natural Science Foundation of China (No. 61475028, No. 61635004, No. 61775023); Science Fund for Distinguished Young Scholars of Chongqing (CSTC2014JCYJJQ40002).
References and links
2. B. Shi, H. Sui, J. Liu, D. Zhang, B. Shi, and H. Sui, “The BOTDR-based distributed monitoring system for slope engineering,” in Proceedings of the 10th IAEG International Congress (2006).
3. A. Klar, I. Dromy, and R. Linker, “Monitoring tunneling induced ground displacements using distributed fiber-optic sensing,” Tunn. Undergr. Space Technol. 40, 141–150 (2014). [CrossRef]
4. K. Lim, L. Wong, W. K. Chiu, and J. Kodikara, “Distributed fiber optic sensors for monitoring pressure and stiffness changes in out-of-round pipes,” Struct. Contr. Health Monit. 23(2), 303–314 (2016). [CrossRef]
5. X. Bao, C. Zhang, W. Li, M. Eisa, S. Elgamal, and B. Benmokrane, “Monitoring the distributed impact wave on a concrete slab due to the traffic based on polarization dependence on stimulated Brillouin scattering,” Smart Mater. Struct. 17(1), 015003 (2008). [CrossRef]
7. X. Angulo-Vinuesa, S. Martin-Lopez, J. Nuno, P. Corredera, J. D. Ania-Castanon, L. Thevenaz, and M. Gonzalez-Herraez, “Raman-Assisted Brillouin Distributed Temperature Sensor Over 100 km Featuring 2 m Resolution and 1.2 °C Uncertainty,” J. Lightwave Technol. 30(8), 1060–1065 (2012). [CrossRef]
9. H. Iribas, A. Loayssa, F. Sauser, M. Llera, and S. Le Floch, “Cyclic coding for Brillouin optical time-domain analyzers using probe dithering,” Opt. Express 25(8), 8787–8800 (2017). [CrossRef] [PubMed]
10. K. Y. Song, Z. He, and K. Hotate, “Distributed strain measurement with millimeter-order spatial resolution based on Brillouin optical correlation domain analysis,” Opt. Lett. 31(17), 2526–2528 (2006). [CrossRef] [PubMed]
11. K. Y. Song and K. Hotate, “Distributed Fiber Strain Sensor With 1-kHz Sampling Rate Based on Brillouin Optical Correlation Domain Analysis,” IEEE Photon. Technol. Lett. 19(23), 1928–1930 (2007). [CrossRef]
13. N. Guo, L. Wang, H. Wu, C. Jin, H.-Y. Tam, and C. Lu, “Enhanced Coherent BOTDA System Without Trace Averaging,” J. Lightwave Technol. 36(4), 871–878 (2018). [CrossRef]
14. H. Wu, L. Wang, Z. Zhao, N. Guo, C. Shu, and C. Lu, “Brillouin optical time domain analyzer sensors assisted by advanced image denoising techniques,” Opt. Express 26(5), 5126–5139 (2018). [CrossRef] [PubMed]
16. A. Voskoboinik, A. E. Willner, and M. Tur, “Extending the Dynamic Range of Sweep-Free Brillouin Optical Time-Domain Analyzer,” J. Lightwave Technol. 33(14), 2978–2985 (2015). [CrossRef]
18. A. Voskoboinik, D. Rogawski, H. Huang, Y. Peled, A. E. Willner, and M. Tur, “Frequency-domain analysis of dynamically applied strain using sweep-free Brillouin time-domain analyzer and sloped-assisted FBG sensing,” Opt. Express 20(26), B581–B586 (2012). [CrossRef] [PubMed]
20. D. Zhou, Y. Dong, B. Wang, C. Pang, D. Ba, H. Zhang, Z. Lu, H. Li, and X. Bao, “Single-shot BOTDA based on an optical chirp chain probe wave for distributed ultra-fast measurement,” Light Sci. Appl. (to be published).
21. Y. Mizuno, N. Hayashi, H. Fukuda, K. Y. Song, and K. Nakamura, “Ultrahigh-speed distributed Brillouin reflectometry,” Light Sci. Appl. 5(12), e16184 (2016). [CrossRef]
23. Y. Dong, D. Ba, T. Jiang, D. Zhou, H. Zhang, C. Zhu, Z. Lu, H. Li, L. Chen, and X. Bao, “High-Spatial-Resolution Fast BOTDA for Dynamic Strain Measurement Based on Differential Double-Pulse and Second-Order Sideband of Modulation,” IEEE Photon. J. 5(3), 2600407 (2013). [CrossRef]
24. I. Sovran, A. Motil, and M. Tur, “Frequency-Scanning BOTDA With Ultimately Fast Acquisition Speed,” IEEE Photon Technol. Lett. 27(13), 1426–1429 (2015). [CrossRef]
25. S. Le Floch, F. Sauser, M. Llera, and E. Rochat, “Novel Brillouin Optical Time-Domain Analyzer for Extreme Sensing Range Using High-Power Flat Frequency-Coded Pump Pulses,” J. Lightwave Technol. 33(12), 2623–2627 (2015). [CrossRef]
28. F. Baronti, A. Lazzeri, R. Roncella, R. Saletti, A. Signorini, M. A. Soto, G. Bolognini, and F. Di Pasquale, “SNR enhancement of Raman-based long-range distributed temperature sensors using cyclic Simplex codes,” Electron. Lett. 46(17), 1221–1223 (2010). [CrossRef]