Open Access
Add to BibSonomyAbstract
We present a new sensing demodulation approach by monitoring the amplitude changes of correlation peak through using optical wideband chaos. For the static strain sensing, the reflection intensity of optical wideband chaos can be modulated by the strain induced wavelength spacing between the wavelength division multiplexing (WDM) device and the sensing grating. Thus, the relative amplitude change (RAC) of correlation peak is mainly determined by the change of chaotic reflection intensity. For the dynamic strain sensing, the reflection intensity of optical wideband chaos can be modulated by the fast fluctuant evanescent wave on a section of no-core fiber (NCF). Thus, the response from correlation peak’s RAC is mainly according to the dynamic strain frequency. The experimental measurements show that a high strain sensing sensitivity of 7.04*10-3 RAC/με is achieved within the measurement range of 900 με during the static strain test. While in the dynamic test, the demodulation can detect the vibration frequency of 6 kHz located at 6 km long. This demodulation method can simultaneously achieve static/dynamic sensing and precisely locating the fiber break point with the high accuracy of several centimeters, making it very easy for network maintenance.
© 2015 Optical Society of America
1. Introduction
Fiber sensor networks usually require a long time to work in long distance, large temperature difference, high pressure, strong magnetic field or harsher natural environment. Thus, the fiber fracture could happen occasionally [1]. It may cost a long time or waste a lot of manpower to find the fault position, since the optical fiber sensing network always has a large and complex coverage area [2]. The traditional demodulation methods used in the optical sensing are usually based on the optical power/wavelength/phase modulations [3]. However, these modulation methods all cannot apparently provide the fiber fault locating ability. Moreover, the various optical time domain reflectometries (OTDRs), such as phase OTDR [4], polarization OTDR [5], have been largely used in the static/dynamic sensing applications. But for the requirement of high locating resolution of ~cm, the generation of complex pulse codes and the strict lasing source with ultra-short bandwidth and low frequency shift have to be adopted [6].
In recent years, the research and application of optical chaos have attracted more and more attention of researchers. The applications of optical chaos have focused on the signal communication, signal processing, signal control and many other fields. For example, the optical chaos have been widely used in the chaos secure optical communication [7,8], the chaotic laser radar ranging [9], the fiber breakpoint detection [10,11], the regulation of laser’s coherent length [12], the generation of real random signal [13], etc.
To apply the optical chaos in the optical sensing area has also been investigated. C. Jáuregui et al. theoretically demonstrated a chaotic fiber-ring resonator for measuring several physical parameters (elongation, attenuation, index of refraction) [14]. Recently, a fiber Bragg grating (FBG) quasi-distributed sensing network with a tunable chaotic fiber laser was theoretically proposed and analyzed [15]. In their work, FBGs with different reflective wavelength are supposed to be interrogated through wavelength sweeping. The simulation results reveal hundreds of multiplexing capabilities and a spatial resolution of up to 1.3 cm. Another work demonstrated that a chaotic laser was amplified to 1.2W to generate Brillouin scattering and set up a distributed OTDR system for temperature measurement [16]. In our previous work, simultaneous and precise fault locating in WDM passive optical network (WDM-PON) was achieved using optical wideband chaos [17]. In this work, we propose a new demodulation method to adopt the optical wideband chaos to achieve the static/dynamic strain sensing applications. The demodulation method is based on the relative amplitude change (RAC) of the correlation peak, which is mainly determined by the variation of the reflection intensity of optical chaos. Therefore, the static/dynamic sensing information and precise locating can be simultaneously achieved based on the RAC and time delay of cross-correlation peaks. Generally speaking, the location of FBGs is mainly distinguished by the different reflective wavelength in the traditional FBG based quasi-distributed sensing network. But in our proposed method, the location of FBGs can be distinguished by the different peak positions along the correlation spectrum, which means that the identical FBGs can be adopted in the quasi-distributed sensing network, which will increase the network multiplexing capabilities. Another advantage is that supposing the fiber fracture happens at the certain place between two FBGs with the interval of up to hundreds of meters, through the proposed method, we can locate the fiber break point with the high accuracy of several centimeters, making it very easy for network maintenance.
2. The generation of optical broadband chaos
The experimental setup is shown in Fig. 1. A semiconductor optical amplifier (SOA) ring structure with an isolator (ISO) acts as the chaotic light source. The SOA has the parameters of central wavelength (1500 nm), optical bandwidth (74 nm), saturation output power (14 mW), and the small signal gain (13 dB). The polarization controller (PC) is used to adjust the polarization of the light into the SOA and the erbium doped fiber amplifier (EDFA) can enhance the power of output chaos. A 99:1 optical coupler (OC2) provides 1% transmission light as the reference signal and 99% transmission light as the detected signal. The circulator assures that the return signal can be detected. Two photon detectors (PDs) with 1 GHz bandwidth are used in the correlation detection.
An optical spectrum analyzer (Yokogawa AQ6370C) and a RF spectrum analyzer (Agilent 40GHz E4447A) are used to observe the chaotic output light. A real-time oscilloscope (OSC) with 12.5GHz bandwidth and 50 GSa/s sampling rate is used to record the reference signal and reflected signal. At last, the data recorded in the OSC are processed by the cross correlation method. The characteristics of our proposed optical broadband chaos are demonstrated in Fig. 2(a) with a high quality of autocorrelation curve in Fig. 2(b) (inset: the time series).
Fig. 2 The characteristics of the broadband chaotic source: (a) the optical spectrum, (b) the autocorrelation curve (inset: the time series).
3. The static strain sensing application
The strain sensing setup is constructed with a 100-GHz-spaced WDM device with 2 channels (CH33, CH34) and the corresponding sensing grating at each branch, as shown in Fig. 3.
As we know when the intensity of optical chaotic reflection changes, the amplitude of peak located at the correlation spectrum will also change after the cross correlation calculation. We use the FBGs to act as the strain sensors, due to the very mature FBG sensing technology currently [18]. However, the induced strain can normally linearly alter the wavelength shift of the FBGs. Therefore, we design a scheme to convert the wavelength shift of grating into the intensity variation of the chaotic reflection.
The reflection spectrum of the corresponding sensing grating is chosen to match the transmission spectrum of WDM, as demonstrated in Fig. 4(a). The 3dB-bandwidth of the gratings is nearly 0.8 nm with the high reflectivity of above 20 dB. In the initial, the reflection intensity will get almost 100% reflection after the combined filtering from the WDM and the grating, if we ignore the insertion loss of WDM. When the strain is applied, the wavelength of the grating will shift towards the longer wavelength. Thus, the wavelength spacing between the WDM and the sensing grating will lead to the decrease of the reflection intensity. The changes of reflection spectra under the strain of 500 με and 800 με are shown in Fig. 4(b) and 4(c), respectively. When the reflection spectrum of the sensing grating has no overlap with the transmission spectrum of WDM, no intensity will be obtained from the reflection. The relationship between the power of reflection and the applied strain has been recorded in Fig. 5, which provides an almost linearly strain response from 0 to 1000 με. The main factors that affect the linear range of static sensing are the reflection spectrum of the sensing FBG and the pass band of the corresponding WDM channel. The linear range reaches maximum when the reflection spectrum of FBG matches precisely with the pass band of the WDM channel. Since the amplitude of correlation peak is also related to the data record conditions, such as sampling rate and the record length, in the experiments, the sampling rate is chosen to 6.25 GSa/s and the record length is fixed to 1M pts in every test. Therefore, the relative amplitude change of the correlation peak is mainly determined by the change of the chaotic reflection intensity.
Fig. 4 The changes of reflection spectra of FBG1 with the applied strains of (a)0 με, (b) 500 με, (c) 800 με.
In Figs. 6(a) and 6(b), the correlation spectra under free of strain and applied strain are drawn. It is seen that the amplitudes of correlation peaks at CH33 and CH34 both decrease on the correlation spectrum Fig. 6(b). We record the correlation spectra under different strain from 0 to 900με in Fig. 7(a). The amplitude changes of the peak at CH34 with the strain alternation are plotted in Fig. 7(b). It is seen that the sensitivity from the linear response at CH34 is about 7.04*10-3 RAC/με with a compelling linearity of adjusted R-square 0.99914. The strain sensing range can be within 900 με. When the applied strain is above 900 με, the reflection intensity becomes very low and makes the response curve nearly flat. This kind of demodulation method can fully take the advantage of the whole information on the correlation spectrum. For example, the sensing position or the fiber fault position can be obtained from the location of peak on the correlation spectrum, as well the sensing amount from the amplitude changes of the correlation peak. From Fig. 6, the distance difference between the two branches is about 114 meters, which fits well with the length of the additional SMF we used at CH33 in the sensing setup.
Fig. 7 (a) The correlation spectra under different strain; (b) The amplitude of correlation peak versus the strain.
4. The dynamic strain sensing application
The schematic of the dynamic sensing setup is shown in Fig. 8. The key component is a section of no-core fiber (NCF) with the length of 29.2 mm, which is spliced to the lead-in single mode fiber (SMF) and lead-out SMF (see inset of Fig. 8). NCF is a special fiber which has the same diameter of 125 μm as SMF, but only contains solid cladding and coating made of pure fused silica materials and polymer materials, respectively. Driven by the periodic sinusoidal signal with a selected microwave frequency, the standard piezoelectric transducer (PZT) continuously vibrates, sequentially applies the vibration force to the sensing structure. The tiny bending of NCF leads to an increase of transmission loss induced by the leaking modes. When NCF vibrates, the light intensity will be modulated by the vibration. In our previous work, SM-NC-SM fiber structure can detect continuous vibration disturbances with the frequencies up to 29 kHz [19].
The modulated intensity waveform of the light going through the vibrated SM-NC-SM fiber structure is monitored and recorded in Fig. 9. The vibration frequency is 24 kHz, and it can be seen that the original chaotic light is modulated with a sinusoidal waveform of the frequency in accordance with the vibration frequency. It further verifies that the intensity modulation can be achieved when the light transmits through on the NCF section. The faraday reflection mirror (FRM) next to the sensing fiber structure can transfer the transmission light signal to the reflection light signal, and transmits through the sensing fiber structure again, thus further enhance the intensity modulation effect induced by the vibration.
Fig. 9 The modulated intensity waveform of the light going through the SM-NC-SM fiber structure vibrating at 24 kHz.
For the locating of sensing point, the reflected signal and the reference signal are recorded simultaneously for a certain period of time, with the OSC. Then the cross correlation is carried out between the two recorded signals. The position of the correlation peak indicates the time delay between the two signals, and provides the location information of the sensing point.
For the application of dynamic sensing, the reflected signal and the reference signal can be divided into segments in time serial and the corresponding segments are processed with cross correlation calculation. The correlation peak amplitude of each segment is determined by the reflection light intensity at different recording time slot, which fluctuates with the vibration. Thus, each correlation peak can be seen as a sampling point of vibration. According to Nyquist sampling law, the sampling rate (SR) should be at least twice as much as the vibration frequency.
Where is the vibration frequency, is the number of segments and is the total sampling time. To keep a correlative relationship between the reflected signal and the reference signal, the time length of a single segment should be larger than the time delay length between the two signals.Where is the sensing length, is the refractive index of SMF core, and c is the speed of the light in the vacuum.Thus, the maximum detection frequency is determined by the sensing length and the sampling number, as the following:
Figure 10(a) demonstrates the correlation curve of 25 segments spliced in time order. Figure 10(b) gives the corresponding envelope of multiple correlation peaks and DC components. It can be seen that both the correlation peaks and the DC components are affected by the vibration, and the demodulated fluctuation frequency here is in agreement with the actually applied vibration frequency of 24 kHz.
Fig. 10 (a) The sampling multiple correlation peak profile, (b) envelope of multiple correlation peaks and DC components.
In order to verify the correctness of formula 4, the demodulation tests under different vibration frequency at different sensing distance are performed. Figures 11(a) and 11(b) demonstrate the vibration of 6 kHz at 6km. The sampling rate is 15.6 kHz, which is selected according to formulas 1 and 3. It is clearly seen that the peak frequency in Fig. 11(b) is just located at 6 kHz through the Fourier transformation. If the vibration locates at 6 km far away, according to formula 4, the maximum detectable frequency is about 8.6 kHz. When the vibration exceeds 8.6 kHz, the system will not provide the detection frequency information of vibraction. We test the vibration frequency of 15 kHz at 6 km and show the experimental results of correlation and Fourier transformation in Figs. 11(c) and 11(d). Since 15 kHz vibration exceeds the maximum detectable frequency (8.6 kHz) at 6 km, the sampling rate cannot be properly chosen to fulfill both formulas 1 and 3. In the test, the sampling rate is set to be 34.4 kHz, which fulfill the Nyquist sampling law, but each segment is too short to construct a correlation peak due to the high sampling rate. Therefore, the vibration frequency cannot be demodulated correctly from Fig. 11(d). If the demodulation of higher frequency vibrations is needed, the sensing distance needs to be shortened. Figures 11(e) and 11(f) show another vibration test of 24 kHz at around 80 m. The frequency peak located on the Fourier transformation in Fig. 11(f) can clearly present the induced vibration of 24 kHz.
Fig. 11 (a) and (b), the spectra of correlation and the demodulated frequency of 6 kHz located on 6 km, respectively, (c) and (d) the spectra of correlation and the demodulated frequency of 15 kHz located on 6 km, respectively, (e) and (f) the spectra of correlation and the demodulated frequency of 24 kHz located on 80 m, respectively.
It is important to choose an appropriate sampling rate when processing the dynamic sensing data. In the first place, the sampling rate should obey formulas 1 and 3. Under such condition, a high sampling rate would slow down the demodulation speed, while a low sampling rate may lower the accuracy of frequency demodulation. Therefore, there needs to be a tradeoff for the selection of sampling rate.
One more advantage of applying the optical chaos to the dynamic sensing area is the ability of fiber fault position locating. If a breaking point occurs in the sensing path, a new correlation peak will appear with steady amplitude. The position of the breaking point can be acquired from the position of the new correlation peak with unchanged relative amplitude.
This system also shows a possibility of multi-point sensing along one fiber line. Different points can be distinguished by different positions of correlation peaks. The vibration frequency at each point can be detected by the amplitude variation of the corresponding correlation peak. Therefore, multi-point vibrations with multiple frequencies can be demodulated at the same time. However, in the direct intensity demodulation, only one vibration point can be demodulated, because it cannot distinguish the fluctuated intensity of the reflected light from two or more points. The wideband chaos also can guarantee the usage of multiple multiplexing methods together in the fiber sensing area, such as time domain multiplexing (TDM) and wavelength domain multiplexing (WDM).
5. Conclusions
In summary, we demonstrate a proof-of-concept experiment for the static/dynamic strain sensing using the optical broadband chaos and a proposed correlation demodulation method. This new demodulation method is not limited in the static/dynamic strain sensing application, but also can be used in temperature, refractive index, acceleration, liquid level, etc, as long as the intensity of chaotic reflection can be modulated by these sensing parameters. Therefore, our scheme will have promising and widespread applications in the simultaneous and precise locating of the sensing positions and the sensing amount among the large number of branches in the sensing networks, as well as monitoring the fiber fault condition at each branch. Therefore, the proposed novel demodulation principle is very suitable for multi-parameter sensing, long coverage distance, high locating resolution of fiber breakpoint, easy networking and convenient maintenance.
Acknowledgment
This work is supported by sub-Project of the Major Program of the National Natural Science Foundation of China (No. 61290315)
References and links
1. R. A. Perez-Herrera, M. Fernandez-Vallejo, and M. Lopez-Amo, “Robust fiber-optic sensor networks,” Photon. Sens. 2(4), 366–380 (2012). [CrossRef]
2. G. R. Kirikera, O. Balogun, and S. Krishnaswamy, “Adaptive fiber Bragg grating sensor network for structural health monitoring: Applications to impact monitoring,” Struct. Health Monit. 10(1), 5–16 (2011). [CrossRef]
3. B. Lee, “Review of the present status of optical fiber sensors,” Opt. Fiber Technol. 9(2), 57–79 (2003). [CrossRef]
4. Y. L. Lu, T. Zhu, L. Chen, and X. Y. Bao, “Distributed vibration sensor based on coherent detection of phase-OTDR,” J. Lightwave Technol. 28(22), 3243–3249 (2010).
5. Z. Zhang and X. Bao, “Distributed optical fiber vibration sensor based on spectrum analysis of Polarization-OTDR system,” Opt. Express 16(14), 10240–10247 (2008). [CrossRef] [PubMed]
6. C. Pan, H. Zhu, B. Yu, Z. Zhu, and X. H. Sun, “Distributed optical-fiber vibration sensing system based on differential detection of differential coherent-OTDR,” in Proceedings of IEEE Conference on Sensors (IEEE, 2012), pp. 1–3. [CrossRef]
7. A. Argyris, D. Syvridis, L. Larger, V. Annovazzi-Lodi, P. Colet, I. Fischer, J. García-Ojalvo, C. R. Mirasso, L. Pesquera, and K. A. Shore, “Chaos-based communications at high bit rates using commercial fibre-optic links,” Nature 438(7066), 343–346 (2005). [CrossRef] [PubMed]
8. C. R. Mirasso, S. Poinsot, L. Larger, M. Peil, and I. Fischer, “Optical chaos communications,” in OPTO-Ireland (International Society for Optics and Photonics, 2005), pp. 139–151.
9. F. Y. Lin and J. M. Liu, “Chaotic radar using nonlinear laser dynamics,” IEEE J. Quantum Electron. 40(6), 815–820 (2004). [CrossRef]
10. A. B. Wang, N. Wang, Y. B. Yang, B. J. Wang, M. J. Zhang, and Y. C. Wang, “Precise fault location in WDM-PON by utilizing wavelength tunable chaotic laser,” J. Lightwave Technol. 30(21), 3420–3426 (2012). [CrossRef]
11. Y. C. Wang, B. J. Wang, and A. B. Wang, “Chaotic correlation optical time domain reflectometer utilizing laser diode,” IEEE Photonics Technol. Lett. 20(19), 1636–1638 (2008). [CrossRef]
12. M. Peil, I. Fischer, W. Elsäßer, S. Bakić, N. Damaschke, C. Tropea, S. Stry, and J. Sacher, “Rainbow refractometry with a tailored incoherent semiconductor laser source,” Appl. Phys. Lett. 89(9), 091106 (2006). [CrossRef]
13. I. Kanter, Y. Aviad, I. Reidler, E. Cohen, and M. Rosenbluh, “An optical ultrafast random bit generator,” Nat. Photonics 4(1), 58–61 (2010). [CrossRef]
14. C. Jáuregui, J. M. López-Higuera, A. Cobo, O. M. Conde, and J. Zubía, “Multiparameter sensor based on a chaotic fiber-ring resonator,” J. Opt. Soc. Am. B 23(10), 2024–2031 (2006). [CrossRef]
15. X. Y. Zhang and L. Z. Yang, “A fiber Bragg grating quasi-distributed sensing network with a wavelength-tunable chaotic fiber laser,” Syst. Sci. Control Eng. 2(1), 268–274 (2014). [CrossRef]
16. Z. Ma, M. J. Zhang, Y. Liu, X. Y. Bao, H. Liu, Y. N. Zhang, and Y. C. Wang, “Incoherent Brillouin optical time-domain reflectometry with random state correlated Brillouin spectrum,” IEEE Photonics J. 7(4), 6100407 (2015). [CrossRef]
17. L. Xia, D. Huang, J. Xu, and D. Liu, “Simultaneous and precise fault locating in WDM-PON by the generation of optical wideband chaos,” Opt. Lett. 38(19), 3762–3764 (2013). [CrossRef] [PubMed]
18. R. Fallon, L. Zhang, A. Gloag, and I. Bennion, “Identical broadband chirped grating interrogation technique for temperature and strain sensing,” Electron. Lett. 33(8), 705–707 (1997). [CrossRef]
19. Y. Ran, L. Xia, Y. Han, W. Li, J. Rohollahnejad, Y. Wen, and D. Liu, “Vibration fiber sensors based on SM-NC-SM fiber structure,” IEEE Photonics J. 7(2), 1–7 (2015). [CrossRef]
