A BOTDA sensing scheme combined frequency sweeping and slope-assisted techniques is proposed and experimentally demonstrated for simultaneously temperature and strain-induced vibration sensing. In this scheme, during sweeping Brillouin gain spectrum (BGS) for temperature measurement, we simultaneously perform FFT to the time-domain traces whose probe-pump frequency difference (PPFD) is within the FWHM of the BGS at each position of fiber, and the location and the frequency of the strain-induced vibration event can be acquired based on SA-BOTDA technique. In this way, the vibration can be continuously measured at each selected working frequency point during the BGS scanning process and multiple measurements of vibration event can be completed in one whole BGS scanning process. Meanwhile, double sidebands probe method is employed to reduce the nonlocal effects. In our experiment, a temperature event and two vibration events with the frequency of 7.00Hz or 10.00Hz are simultaneously measured near the end of 10.6km long sensing fiber in a traditional BOTDA system. The system shows 1.2°C temperature accuracy and 0.67Hz frequency resolution, as well as a 3m spatial resolution. The proposed method may find some potential applications where both the strain-induced vibration frequency and temperature are the diagnostic objects.
© 2016 Optical Society of America
Brillouin optical time domain analyzer (BOTDA) sensors have been widely used in large civil engineering due to their capability to monitor various environmental parameters distributions, such as temperature and strain, along a long distance [1–3]. The sensing mechanism of classical BOTDA system is based on the stimulated Brillouin scattering (SBS) of optical fiber, in which two counter-propagating lightwaves, a continuous-wave (CW) probe signal and a pump pulse, interact with each other though the acoustic wave. When the probe-pump frequency difference (PPFD) is within the fiber Brillouin gain spectrum (BGS), power transfer between the two optical signals. If the PPFD equals to the Brillouin frequency shift (BFS) of the fiber, the transferred power reach a maximum value. Therefore, the BFS can be measured by sweeping the Brillouin gain spectrum. And thus the distributed strain or temperature sensing can be realized by employing the dependence of the BFS on strain or temperature .
In classical BOTDA, the optical frequency of either the pump or probe waves is swept across 100-200MHz to recover the BGS along the sensing fiber, which together with the need for averaging to improve SNR, result in a fairly slow procedure, limit classical BOTDA method to static parameter measurements, such as temperature and strain. In order to extend the BOTDA to dynamic parameter sensing, many approaches such as the correlation-domain method (BOCDA) [5,6], the sweep-free distributed Brillouin time-domain analyzer (SF-BOTDA) method [7,8], Brillouin phase-shift measurement method  and slope-assisted BOTDA technique (SA-BOTDA) [10–13] have been proposed. The BOCDA technology can realize a high spatial resolution in dynamic strain sensing. But this method is an essentially discrete sensing technology. It cannot acquire information at all the locations along the sensing fiber. In SF-BOTDA, multiple pumps and multiple probes are employed to simultaneously sample the time domain traces at many frequency points to reconstruct the BGS. Recently, a modified SF-BOTDA using two sets of pump-probe lights generated by two individual laser sources have also been proposed and experimentally demonstrated . The measurement speed is much faster than convectional BOTDA, but its implementation is very sophisticated, requiring a precise control of multiple probe-pump frequency difference. The Brillouin phase-shift method could achieve a high measurement precision and effectively reduced the pump power dependence, as well as a large dynamic range but at the expense of measurement complexity. In 2013, a high-spatial-resolution fast Brillouin optical time-domain analysis (BOTDA) for distributed dynamic strain measurement based on differential double-pulse and second-order sideband of modulation was demonstrated in a sensing distance of 50m Panda fiber. The system shows a high spatial resolution of 20 cm and the strain accuracy of 14με but a relatively short sensing length . In 2015, an ultimate possible measurement speed of frequency scanning BOTDA employing fast frequency tuning and polarization diversity was proposed and experimentally demonstrated. The obvious merit of this presented sensor is that no averaging is needed and the measurement speed is limited only by the number of scanning frequencies and the fiber length . However, this method also shows a relatively complicated control and measurement process. Most recently, a multi-slope assisted fast Brillouin optical time-domain analysis (F-BOTDA) is proposed and demonstrated. The system exhibits a maximum strain variation up to 5000με in a 32m PM fiber with a spatial resolution of ~1m . However, it seems that the timing and control are relatively complex.
Although the dynamic range of strain-induced vibration in SA-BOTDA is subject to the spectrum width of Brillouin Gain Spectrum (BGS) and the frequency range of vibration is limited by the sensing distance, it is the most promising and effective technology for dynamic measurement in practical field application. In SA-BOTDA, the PPFD is commonly fixed at the middle of either the falling or rising slopes of the BGS. The vibration-induced modulation of the local BFS is measured as an intensity variation of the probe wave along the sensing fiber. Strain information then is computed according to the measured intensity variation. Most recently, a modification of SA-BOTDA, called double-slope-assisted BOTDA (DSA-BOTDA), was proposed to eliminate the pump pulse power dependence, and the demonstrated system can efficiently keep the strain measurements away from the power variations of pump pulse .
Considering the need of frequency sweeping to map the BGS for temperature or strain measurement in convectional BOTDA and only one working frequency point is required for measurement in SA-BOTDA, we propose a simultaneous temperature and strain-induced vibration measurements scheme in classical BOTDA system by combining frequency sweeping and slope-assisted techniques. In our scheme, during scanning BGS process for temperature measurement, the time-domain traces whose PPFD are within the FWHM of BGS are simultaneously selected to realize the strain-induced vibration sensing based on SA-BOTDA technique. In this way, the vibration can be continuously measured in one whole BGS scanning process. Using the proposed method, a temperature event and two vibration events are simultaneously measured near the end of 10.6km long sensing fiber with 3m spatial resolution in our experiment. The obtained temperature accuracy and frequency resolution are ± 1.2 °C and ± 0.67 Hz, respectively.
2. Principle and measurement method
2.1 Measurement process in convectional BOTDA
The measurement process of convectional BOTDA is shown in Fig. 1 (a). Assuming the m scanning probe-pump frequency difference is employed to map the BGS along the sensing fiber. For each PPFD, N averages are needed to enhance the signal-to-noise ratio (SNR). In the practical experiment, successive N optical pulses with the same period T are injected into the sensing fiber at each working frequency point. So N time-domain traces are obtained at each PPDF, where represents the distance between the input end and the scattering point, c is the speed of light in vacuum and n is the refractive index of sensing fiber. It must be pointed out that the period T should be larger than the roundtrip time of the pulse in fiber. By averaging the N time-domain traces, a probe light power distribution curve at PPFD can be obtained. Repeated the above step for m frequency differences, and m probe light power distribution curves can be obtained.
2.2 Dynamic strain measurement in SA-BOTDA
Essentially, dynamic strain measurement in SA-BOTDA is based on the fact that the strain-induced modulation of the local BFS changes the local Brillouin gain of probe wave. The intensity variation of the probe light along the sensing fiber reflects the variation of local BFS, and thus the strain information , [12,13], .
In a dynamic strain environment, the BFS at certain location z is a function of both time and location along the fiber, and it can be written as:10]. That is to say, by performing FFT to the time-domain traces whose PPFD is within the FWHM of the BGS, the strain-induced vibration can be located and its frequency can be measured.
2.3 Simultaneous strain-induced vibration and temperature measurements
According to the measuring process of temperature or strain in convectional BOTDA and the vibration sensing in SA-BOTDA described above, SA-BOTDA and convectional BOTDA actually use the same experimental configuration. Therefore, simultaneous temperature and strain-induced vibration measurements may be realized in the same apparatus by combined frequency sweeping and slope-assisted methods.
The concrete scheme is shown in Fig. 1(b). Simultaneously, the N time-domain traces whose PPFD is within the FWHM of BGS are selected during the sweeping frequency process for temperature measurement. Aligning the selected N time-domain traces in time sequence, and performing FFT to the selected time-domain traces at each position of fiber, the frequency spectrum including the vibration information can be acquired. Therefore, the position and frequency of the vibration could be measured. A disadvantage of this vibration sensing scheme is that its vibration frequency measurement range is limited by the spectrum width of Brillouin Gain Spectrum (BGS) and the sensing distance, which is a common shortcoming of the SA-BOTDA. It should be noted that the averaged effect of vibration on the time-domain trace tends to zero after N averages, which is because the averaged time-domain trace contain many vibration periods and the transient contribution of the vibration to the temperature measurement is averaged out . Thus the reconstructed BGS has remained unchanging and temperature measurement can still be realized. In this way, the strain-induced vibration can be continuously measured at each selected working frequency point, but a complete temperature measurement needs scanning the whole BGS.
The vibration events are loaded by two motor-driven-elliptical cams in our experiment, the magnitude of the strain induced by vibration is not constant value. In addition, all the time-domain traces whose PPFD are within the FWHM of BGS can be are selected to perform the vibration measurement, and the amplitudes of each selected time-domain traces is different and hence, it cannot determine the magnitudes of the strain according to the amplitudes of the time-domain traces. Therefore, we did not perform the magnitude of the strain measurement but focused on the vibration frequency measurement.
3. Experimental setup
The experiment setup is shown in Fig. 2. A tunable laser source (TLS) with 100 kHz linewidth and 16-dBm output power operating at 1550.00 nm is employed in experiment. A high extinction ration (ER) modulator consisting of an electro-optic modulator (EOM1) and an optical switch (OS) is applied to convert the 90% branch into optical pump pulse. The OS used here is to suppress the noise created by the leakage light to reach a high ER. The ER in our experiment is over 65dB. Using the modulator, pump pulses with the duration of 30 ns and period of 150us are generated. Then an Erbium doped fiber amplifier (EDFA) is used to amplify the pump pulses to desired power. In order to reduce the polarization-induced influence on SBS, a polarization scrambler (PS) is inserted between the EDFA and the CIR. Finally, the pump pulses are inputted into the fiber under test (FUT) through an optical circulator (CIR). In the 10% branch, two probe sidebands (Stokes and anti-Stokes line) are generated to effectively suppress the nonlocal effect by an electro-optic modulator (EOM2) . The EOM2 is driven by a microwave generator and the working frequency point is adjusted though changing the output frequency of the microwave generator. In order to avoid the interference of pump pulses to the EOM2, an isolator (ISO) is inserted between the EOM2 and the FUT. After propagating through the FUT, the unwanted sideband is filtered by a fiber Bragg grating (FBG) filter with the FWHM of 0.1nm. Then only one probe sideband is detected by the 350MHz photon detector (PD). The electrical signals output from the PD are collected by a digital oscilloscope (Agilent, DSO9254A) and the collected data are then sent to a computer (PC) for further processing. In our experiment, the Stoke sideband is selected, which is corresponding to the Brillouin-gain process in conventional BOTDA.
In our setup, the 10.6 km long sensing fiber contained two spools, the lengths of which are 8.75 km and 1.85 km, respectively. The FUT layout is shown in Fig. 2(b). Near the end of the last spool 1.85 km fiber, a section of 13 m is set in an oven for temperature control. Two vibration events spacing 28 m are generated by two motor-driven-elliptical cams. The first vibration event is 15 m away from the oven and the lengths of the fibers stretched on the cams are 9 m and 7 m, respectively. In initial, the two segments fiber attached on the cams are loose. It must be pointed out that the vibration will be loaded twice a round on the fiber.
4. Results and discussion
Firstly, we got the preliminary BGS along the sensing fiber by sweeping the pump-probe frequency difference over 106MHz, from 10.775 to 10.891GHz, with the scanning step of 2MHz. It should be noted that the sweeping frequency range is large in our experiment, which is because there are 24 MHz difference in the BFS values of the two spools fiber used. The sweeping frequency range can be further reduced if the fiber with a uniform BFS is used in practical application, and thus lesser measurement time is needed. The time-domain traces along the fiber at each frequency point were collected after 4000 averages with 200 MSa/s sampling rate (corresponding to 0.5m sampling interval). The BFS distribution along the fiber can be obtained via fitting the measured BGS with Lorentzian curves. The measured average BFSs of the two spools are 10.817GHz and 10.841GHz, respectively. And the measured FWHM of their BGS is uniform 46MHz. Clearly, the time-domain traces of that the PPFD ranges from 10.818GHz to 10.864GHz can be selected to realize the vibration sensing in the second spool fiber by using the SA-BOTDA method.
4.1 Strain-induced vibration measurement
In order to confirm that the time-domain traces of the PPFD within the FWHM of BGS can be used for strain-induced vibration measurement, the time-domain traces of the frequency points when the PPFD is respectively set at 10.815GHz, 10.831GHz, 10.845GHz and 10.865GHz are selected to perform the vibration measurement. It is obvious that the selected frequency points are within or near the FWHM of BGS of the second spool fiber. In this experiment, the rotating speeds of the two elliptical cams are uniformly set to 310 rounds per minute (RPS), corresponding to a variation frequency of 10.33Hz. The period of the pulse is 150us and 10 times averaging is applied to collect each time-domain trace for reaching a high SNR. For each frequency point (i.e. each PPFD), one complete measurement lasted for 1.5s, corresponding to 1000 collected time-domain traces. Considering the memory deep of the OSC, we only collect the last 50us signal in each pulse period with 200 MSa/s sampling rate. The vibration frequency can be obtained by implementing FFT to the 1000 time-domain traces . According to the Nyquist sampling theorem, the maximum measurable frequency is 333.3Hz and the frequency resolution is 0.67 Hz in our case. The measurable frequency range is suitable for distributed monitoring of vibration of large infrastructure, including bridges, tunnels, rock slopes pipelines . Three-dimensional vibration frequency spectra ranging from10.37km to10.65km can be obtained by performing FFT to each position along the sensing fiber after aligning the selected 1000 time-domain traces in time sequence. The results are shown in Fig. 3. It can be seen that the two loaded variation events, locating at 10.570km and 10.615km, respectively, can be successfully identified, and the measured variation frequencies are uniform 10.00Hz, which agree well with the real vibration frequency (10.33Hz) loaded by the two cams. Although the second harmonic generation is also observed in Fig. 3(d), the vibration frequency can still be distinguished according to the fundamental frequency.
In order to further demonstrate the sensor’s ability for measuring different frequency vibrations along the same sensing fiber, we perform the similar measurements as above when the rotating speed of the first elliptical cam is sequentially set at 310 RPM and 210 RPM, respectively. For comparison, the time traces of the first vibration point are plotted in Fig. 4 when the PPFD points are respectively set at 10.815GHz, 10.831GHz and 10.845GHz. In this figure, the blue lines and the green lines are the time-domain traces when the speeds of the cam are 310 RPS and 210 RPS, respectively. It is clear that the periods of the blue lines and the green lines in the Figs. 4(a)-4(c) are respectively 0.10s and 0.14s, corresponding to 10.00Hz and 7.00Hz, which are in agreement with the rotation speed of the cam. To determine the accurate vibration frequency, the traces for the PPFD point at 10.831GHz are analyzed by using FFT method, the frequency domain spectra are shown in Fig. 4(d), and they display the fundamental frequency of 10.00Hz and 7.00Hz. It can be observed from both the time and frequency domain spectra that the vibration frequencies can be clearly identified.
4.2 Simultaneous multi-parameters measurement
In the experiment, a 13-m-long fiber in an oven is heated to 45 °C, while the rest of the fiber is keeping at ambient temperature of 30 °C. At the same time, the vibration events are applied on the fiber. The rotating speeds of the two elliptical cams are about 310 RPS, corresponds to 10.33Hz vibration frequency. The data acquisition and processing are the same as description in section 2.3. The period of the pulse in this experiment is 150us. The BGS is reconstructed by sweeping the PPFD from 10.800 to 10.900GHz with the step of 2MHz. thus 51 probe light power distribution curves(m = 51) are collected. In order to increase SNR, The time-domain traces at each working point (PPFD) average 1000 times (j = 1000), corresponding to measurement time of 1.5s and hence, a complete scanning time needs 76.5s. During the scanning process, each 10 time-domain traces are averaged as a time-domain trace at each working point, and there are 100 averaged time-domain traces are collected to perform FFT for vibration measurement. Obviously, a complete vibration measurement lasts 1.5s, and 51 measurements of vibration event can be completed in one whole BGS scanning process in theory. In our experiment, just the averaged time-domain traces whose PPFD are within the FWHM of BGS are simultaneously selected to perform the vibration measurement during the scanning frequency process.
The measured three-dimensional BGS along the fiber and the BGS at specific locations are shown in Fig. 5. It is visible from Fig. 5(a) that the BGS of the heated segment is well distinguished from the rest fiber. The spectra shown in Fig. 5(b) are corresponding to the locations at 8.50km, 10.00km and 10.57km, respectively. The three BGSs are undistorted, which confirm that nonlocal effects can be negligible in our experiment due to employment of the two sidebands probe method . The measured FWHMs of the Brillouin gain spectra at three locations are uniform 46MHz.
The BFS distribution along the fiber can be obtained and is shown in Fig. 6. The temperature accuracy can be estimated from the standard deviations of the measured BFS . In our experiment, we obtained the standard deviations of ± 1.0 MHz and ± 1.2 MHz for the two spools sensing fiber (not including the heated segment), which give the temperature accuracy of ± 1.0 °C and ± 1.2 °C, respectively. The mean values of BFS for the first and the second segments are 10.8175 GHz and 10.8415 GHz, respectively. The frequency difference between the heated and the rest of the second spool fiber is approximately 15.1 MHz, which gives a temperature difference of 15.1 °C using the sensitivity of 1.0 MHz/°C, agreeing well with the expected value of 15.0°C. The achieved spatial resolution as defined in  is about 3 m as shown in the inset of Fig. 6.
The time-domain traces at the working frequency point ranging from 10.814GHz to 10.866GHz are selected to perform the vibration measurement. The typical measured results respecting the information of vibration events ranging from 10.37km to 10.65km at the PPFD of 10.831GHz are shown in Fig. 7. The 3D vibration frequency spectrum is shown in Fig. 7 (a) and the inset in Fig. 7(b) is the time-trace near the first cam. The two vibration events can be obviously diagnosed according to the time-domain trace, and the achieved spatial resolution estimated from the inset is about 3-m, which is in good agreement with the use of 30 ns probe pulse.
Since only the time-domain traces whose working frequency point is within the FWHM of BGS are selected to perform the vibration measurement, some vibration events may be missed when the sweeping frequency point is out of the FWHM of BGS. In other word, there is a time dead-zone in our system, which is a restraint in our system. In fact, the time dead-zone can be reduced through narrowing the sweeping frequency range. But in this way, it may reduce the temperature measuring range. So it is a trade-off between the time dead-zone and temperature measuring range in practical application and deserves further research.
In conclusion, we have confirmed that all the time-domain traces at the frequency point when the PPFD is within the FWHM of BGS can be used for vibration measurement based on the slope-assisted method. Thus simultaneous strain-induced vibration and temperature measurement could be realized by combined the frequency sweeping and slope-assisted techniques. We have demonstrated a temperature event with 15.1 °C above ambient temperature and two vibrations with frequency of 7.00Hz or 10.00Hz are simultaneously measured near the end of 10.6km long sensing fiber in a traditional BOTDA system. The obtained temperature accuracy and frequency resolution are ± 1.2°C and ± 0.67Hz, respectively. Although the dynamic strain amplitude cannot be achieved according to the probe power variation in our present experiment, the proposed method may find some potential applications where both the vibration frequency information and temperature are the diagnostic objects.
This work is supported the National Natural Science Foundation of China under grant Nos. 61307096, 61565002, 61540017 and 61405090, and also is supported by Natural Science Foundation of Guangxi province under Grant No. 2014GXNSFBA118282, and the key project of Guangxi province Higher Educational Science and Technology under Grant No. ZD20140213.
References and links
2. L. Thévenaz, “Inelastic Scatterings and Applications to Distributed Sensing,” in Advanced Fiber Optics: Concepts and Technology (EPFL Press, 2011).
3. X. Angulo-Vinuesa, S. Martin-Lopez, P. Corredera, and M. Gonzalez-Herraez, “Raman-assisted Brillouin optical time-domain analysis with sub-meter resolution over 100 km,” Opt. Express 20(11), 12147–12154 (2012). [CrossRef] [PubMed]
5. K. Y. Song and K. Hotate, “Distributed Fiber Strain Sensor With 1-kHz Sampling Rate Based on Brillouin Optical Correlation Domain Analysis,” IEEE Photonics Technol. Lett. 19(23), 1928–1930 (2007). [CrossRef]
6. K. Y. Song, M. Kishi, Z. He, and K. Hotate, “High-repetition-rate distributed Brillouin sensor based on optical correlation-domain analysis with differential frequency modulation,” Opt. Lett. 36(11), 2062–2064 (2011). [CrossRef] [PubMed]
7. A. Voskoboinik, J. Wang, B. Shamee, S. R. Nuccio, L. Zhang, M. Chitgarha, A. E. Willner, and M. Tur, “SBS-based fiber optical sensing using frequency-domain simultaneous tone interrogation,” IEEE J. Lightwave Technol. 29(11), 1729–1735 (2011). [CrossRef]
11. Q. Cui, S. Pamukcu, W. Xiao, and M. Pervizpour, “Truly distributed fiber vibration sensor using pulse base BOTDA with wide dynamic range,” IEEE Photonics Technol. Lett. 23(24), 1887–1889 (2011). [CrossRef]
12. Y. Peled, A. Motil, L. Yaron, and M. Tur, “Slope-assisted fast distributed sensing in optical fibers with arbitrary Brillouin profile,” Opt. Express 19(21), 19845–19854 (2011). [CrossRef] [PubMed]
14. K. Tsuji, H. Noda, and N. Onodera, “Sweep-free Brillouin optical time domain analysis using two individual laser sources,” Opt. Rev. 19(6), 381–387 (2012). [CrossRef]
15. 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 Photonics J. 5(3), 2600407 (2013). [CrossRef]
16. I. Sovran, A. Motil, and M. Tur, “Frequency-scanning BOTDA with ultimately fast acquisition speed,” IEEE Photonics Technol. Lett. 27(13), 1426–1429 (2015). [CrossRef]
17. D. Ba, B. Wang, D. Zhou, M. Yin, Y. Dong, H. Li, Z. Lu, and Z. Fan, “Distributed measurement of dynamic strain based on multi-slope assisted fast BOTDA,” Opt. Express 24(9), 9781–9793 (2016). [CrossRef] [PubMed]
18. A. Motil, O. Danon, Y. Peled, and M. Tur, “Pump-power-independent Double Slope-Assisted distributed and fast Brillouin fiber-optic sensor,” IEEE Photonics Technol. Lett. 26(8), 797–800 (2014). [CrossRef]
19. L. Zhou, F. Wang, X. Wang, Y. Pan, Z. Sun, J. Hua, and X. Zhang, “Distributed Strain and Vibration Sensing System Based on Phase-Sensitive OTDR,” IEEE Photonics Technol. Lett. 27(17), 1884–1886 (2015). [CrossRef]