Open Access
Add to BibSonomy
Abstract
A digital optical phase-locked loop (OPLL) has been implemented to develop a distributed Brillouin sensing system in optical fibers. In our experiment, two commercial semiconductor lasers are phase-locked to each other with a highly flexible offset frequency using field programmable gate array (FPGA)-based electronics. Then, the difference frequency between the two lasers is highly stabilized and scanned by a desired step frequency in the vicinity of the Brillouin frequency of standard single-mode optical fibers. Consequently, the distribution of Brillouin frequency shift over a 50 km-long sensing fiber has been successfully measured by a very simple and low-cost Brillouin optical time-domain reflectometry (BOTDR) sensing system without any penalty in the sensing performance. The measurement repeatability at 50 km position of sensing fiber with a 5 m spatial resolution was measured be 4.5 MHz under fast measurement conditions: the number of trace averaging of 2000 and the frequency scan step of 12.8 MHz, showing the figure-of-merit of 3.0.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Over the last few decades, distributed optical fiber sensing (DOFS) systems have been extensively developed thanks to their unmatched advantages in terms of long-range sensing capability at high performance [1–3]. Today, DOFS systems are frequently deployed for various industrial applications, i.e., security monitoring of oil & gas pipelines, submarine power cables and the structural health of civil infrastructure. Various approaches have been devised, using different physical mechanisms of light scattering such as Rayleigh, Raman and Brillouin as well as fiber Bragg grating. Among them, Brillouin scattering-based DOFS systems have attracted a significant interest due to their promising performance and are today most-widely implemented for the applications of distributed temperature and strain monitoring [3–5].
Brillouin scattering in optical fibers results from the photon-phonon interaction between incident light and a thermally-excited acoustic wave. Then, the generated acoustic wave acts as a weak fiber Bragg grating that diffracts the incident light. Since the acoustic wave propagates at an acoustic velocity inside optical fibers, the diffracted light experiences an optical frequency shift due to the Doppler effect, referred to as Brillouin frequency shift νB, expressed as the following equation:
where n is the refractive index of the optical fiber, Va is the acoustic velocity, c is the light velocity in vacuum and νP is the optical frequency of the incident light. The acoustic velocity varies monotonically with respect to the amount of physical perturbation (strain and/or temperature) that occurs along the sensing fiber. As a result, the Brillouin frequency tends to have a linear response to the change in temperature and strain. It is the principle of Brillouin-based DOFS systems to perform the quantitative measurement of temperature and strain via the Brillouin frequency.Stimulated Brillouin scattering (SBS) is a nonlinear optical parametric process between two counter-propagating optical waves (pump and probe signals) mediated by an acoustic wave. When the optical frequency difference between the two signals is set to be equal to the acoustic wave frequency, the amplitude of the acoustic wave is reinforced through the electrostriction phenomenon. As a result, the probe signal can experience either a significant optical gain or loss, depending on the phase matching condition. Due to the intrinsic lifetime of the acoustic waves in optical fibers on the order of 10 ns, the spectral bandwidth of the optical gain/loss is as narrow as a few tens of MHz. For this reason, a high stability of the relative frequency between the pump and probe is crucial to obtain a high signal-to-noise ratio; hence, resulting in a high resolution of temperature and strain sensing.
In early development of Brillouin optical time-domain analysis (BOTDA) sensing systems, two distinct free-running lasers were utilized. The two lasers were temperature-tuned until the differential optical frequency was present around the Brillouin frequency, and then the optical frequency of one of the two lasers was fine-tuned by adjusting the injection current [6]. In turn, the Brillouin gain spectrum was mapped in frequency while measuring the heterodyne signal frequency using an electrical spectrum analyzer, which makes the sensing system complex. Soon after, a new method for controlling the relative optical frequency was proposed, using a single laser and an external intensity modulator [7,8]. The laser was first split into two branches to generate pump and probe signals. One branch was then fast-optically gated to generate a pulse train as pump signal. The other branch contained an intensity electro-optic modulator. The light was intensity-modulated at an RF frequency (typically around 10 GHz) to generate two sidebands at Brillouin frequency with a suppressed carrier. Therefore, the distribution of Brillouin gain was acquired as a function of the RF frequency applied to the modulator. However, this scheme requires expensive high bandwidth RF electronics and a modulator. Alternatively, optical phase locking techniques based on injection locking [9] and offset locking [10] were investigated, but the frequency tuning mechanism of those systems was neither convenient nor stable.
In this paper, we propose a novel method to stabilize the relative optical frequency between pump and probe waves, based on a digital optical phase-locked loop (OPLL) using a cost-effective field programmable gate array (FPGA)-based electronics platform. Two optical signals were generated from two distinct commercial semiconductor lasers, and one laser was phase-locked to the other with an offset frequency close to the Brillouin frequency of the sensing fiber. Then, the offset frequency was accurately and conveniently adjusted at a desired frequency, simply by varying the frequency of the numerically-controlled oscillator generated implemented in the FPGA. In general, the double-sideband configuration [8] is preferred to construct a BOTDA sensing system while the proposed configuration is compatible with the single-sideband BOTDA sensing system unless a third laser is employed to generate a second probe signal. So, for the proof-of-concept demonstration we decided to apply this technique to a Brillouin optical time-domain reflectometry sensing system; hence, utilizing the probe signal as an optical local oscillator in the heterodyne detection system. Yet bear in mind that the principle of this technique is completely applicable to realize BOTDA systems.
2. Development of digital OPLL between two semiconductor lasers
Figure 1 depicts the simplified diagram of our digital OPLL, in which the slave laser is optically phase-locked to the master laser. The temperature and current of two semiconductor lasers are individually controlled, so that they are spectrally positioned with a desired offset frequency, i.e., in a range of 10 to 12 GHz for the Brillouin sensing purpose of our experiment. The outputs of the two lasers are combined via an optical coupler and, when the combined light illuminates the photodetector, an RF heterodyne signal is generated at frequency fS that is the difference frequency between the two lasers. In conventional OPLLs, the heterodyne signal is mixed with a stable RF local oscillator at frequency fLO to compare the phase difference. Then, the optical frequency of the slave laser is modified by feeding an error signal to the slave laser controller; hence, stabilizing the phase difference. As a result, the optical frequency of the slave laser becomes phase-locked to the master laser frequency, meaning that the difference frequency between master and slave lasers remains stable at fLO, referred to as the offset frequency. Moreover, the offset frequency is flexibly adjustable, simply by changing the RF frequency of the local oscillator since the heterodyne signal frequency will follow the local oscillator within a certain feedback bandwidth. Then, to build up a Brillouin sensing system the master laser is utilized to generate a Brillouin pump signal while the slave laser plays the role of either the Brillouin probe signal for BOTDA systems or the optical local oscillator for BOTDR systems.
Fig. 1. Schematic diagram of digital OPLL based on FPGA electronics platform. FPGA: field programmable gate array, PLL: phase-locked loop, OC: optical coupler. Black lines: electrical lines and pink lines: optical line.
In our digital OPLL scheme, a frequency divider is implemented to scale down the bandwidth and the cost of necessary electronics without any significant degradation of the PLL performance [11]. This way the phase detection process is considerably simplified thanks to the frequency down-conversion. In our experiment, the heterodyne signal frequency is divided by a frequency division ratio of n = 256. Consequently, the frequency-divided heterodyne signal is down-converted from ∼10 GHz to a few tens of MHz. It allows us to realize a fully digital PLL with the phase detector, the local oscillator, the loop filter and the feedback gain, which are all implemented into the FPGA. Here, a cost-effective compact FPGA board, STEMlab 125-14 from Red Pitaya platform is utilized. The digital part of the OPLL design as described in Fig. 2 is implemented using the oscimpDigital ecosystem that provides open-source tools for RF applications [12].
Fig. 2. Stabilization design implemented in the SoC FPGA. NOC: numerically controlled oscillator, FIR: finite impulse response, ADC: analog-to-digital converter, DAC: digital-to-analog converter and PI: proportional and integral.
In the digital OPLL circuit, the frequency-divided signal is demodulated in frequency, using a numerically controlled oscillator (NCO), and filtered by a finite impulse response lowpass filter. The resulting baseband signal corresponds to the phase error signal of the OPLL. This error signal is then sent to a proportional-integral (PI) controller to generate a corresponding correction signal. Finally, the correction signal is delivered to the slave laser driver for feedback on the laser current. The adjustment of the parameters in our FPGA design such as the demodulation frequency, the FIR filter, and the proportional and integral gains of the controller is performed through a web server hosted by the system on chip (SoC). Both the error and correction signals continuously extracted from the FPGA are utilized to optimize the operation parameters and to monitor the PLL status in real-time.
Figure 3(a) depicts the measured phase noise spectrum of the original and frequency-divided heterodyne signals when the slave laser is phase-locked to the master laser with an offset of 10.24 GHz; hence, setting fLO to 40 MHz. The measured phase noise level of the 10.24 GHz tone is measured to be −38 dBc/Hz at ∼270 kHz offset that corresponds to the bandwidth of our digital PLL and it decreases to as low as −55 dBc/Hz for lower offset frequencies. The PLL bandwidth is essentially determined by the characteristics of the loop filter, especially the loop delay generated by the optical path length and the response time of the digital part. [11,13].
Fig. 3. (left) Measured phase noise of the original and frequency-divided heterodyne signal. (right) Measured RF spectrum of the original heterodyne signal at 10.24 GHz with RBW 10 Hz, showing nearly monochromatic beat note.
The SNR defined as a ratio of the carrier power to the spectral density of the noise power is measured to be ∼48 dB higher for the frequency divided signal compared to the original signal. This is in line with the used division factor n = 256 (20*log(256) = 48 dB). Figure 3(b) shows the RF power spectrum of the locked-OPLL output signal, measured by an electrical spectrum analyzer with a resolution bandwidth of 10 Hz. The carrier has a coherent peak with a power 47 dB above the noise pedestal, which is sufficient for a stable SBS interaction.
3. Digital OPLL-based BOTDR sensing system
Figure 4 illustrates the simplified schematic diagram of the proposed Brillouin optical time-domain reflectometry (BOTDR), based on the digital OPLL. Two commercial distributed feedback (DFB) laser diodes are utilized to realize a distributed temperature sensing system, using spontaneous Brillouin scattering in optical fibers. The lasers are operated at 1560 nm with a free-running linewidth of 200 kHz and their optical frequencies are pre-adjusted to be close to each other by controlling the laser TEC temperature and injection current. A standard single mode optical fiber (SMF) with ∼50 km length is used as Brillouin sensing fiber.
Fig. 4. Simplified schematic diagram of Brillouin optical time-domain reflectometry, based on a digital OPLL between two semiconductor lasers. SCR: polarization scrambler. OC: optical coupler. SOA: semiconductor optical amplifier. EDFA: erbium-doped fiber amplifier.
A part of optical power from both lasers are combined to phase-lock the slave laser to the master laser with an offset frequency of 10.8 GHz, which is close to the Brillouin frequency of the sensing fiber. Another branch of the master laser is then fast-optically gated through a semiconductor optical amplifier (SOA); hence, generating an optical pulse train at a repetition rate of 1.9 kHz with a 50 ns duration, corresponding to a 5 m spatial resolution. The optical pulse exiting the SOA is amplified by an erbium-doped fiber amplifier (EDFA), so that the pulse peak power is set to ∼20 dBm, which is just below the typical threshold of the modulation instability in standard SMFs. After passing through a polarization scrambler, the pulsed pump signal is directed to the sensing fiber via an optical circulator. Notice that we use all polarization maintaining fibers (PMF) to build up the optical circuit of the OPLL; hence, securing the stability of the OPLL performance from the detrimental polarization fading effect.
To reconstruct the distribution of the Brillouin frequency shift along the sensing fiber, the optical heterodyne technique developed in the previous work [14] is utilized for our experiment. Therefore, the slave laser plays the role of an optical local oscillator (LO). The spontaneous Brillouin back-scattered signal, which was continuously generated while the pump propagates through the sensing fiber, is directly combined with the LO. So, the mutual interference manifested a heterodyne beat signal at the differential optical frequency between them. The spectral profile of the electrical heterodyne signal is given by the convolution of the LO optical spectrum and the spectrum of the spontaneous Brillouin scattering. However, under conditions that the linewidth of the LO laser is 200 kHz while the spectral width of the Brillouin scattering is typically few tens of MHz, the electrical spectral profile of the heterodyne signal can be approximated to a duplicate of the actual Brillouin spectrum. To develop a simple detection system, a sensitive balanced detector with an electrical bandwidth of 100 MHz is used, and an RF low-pass filter with a cutoff frequency of 20 MHz (corresponding to 5-m spatial resolution), followed by an RF power detector is placed at the detector output for the spectral analysis.
The distribution of spontaneous Brillouin scattering spectrum along the sensing fiber is then acquired while scanning the NCO frequency from 41.2 MHz to 43.2 MHz by steps of 50 kHz, which effectively corresponds to the optical frequency scan of 10.5472 GHz to 11.0952 GHz by a step of 12.8 MHz. Notice that the minimum achievable increment of the NCO frequency generated by the used FPGA is 0.000114 Hz, so the minimum optical frequency scan step can be set to 0.029 Hz, which is much smaller than the typical frequency scan step of 1 MHz in Brillouin sensing systems. Consequently, a 3D map of the distributed spontaneous Brillouin scattering spectrum along the sensing fiber is obtained as shown in Fig. 5(a), and the Brillouin frequency shift and spontaneous Brillouin intensity profile along the sensing fiber are computed from the standard parabolic fitting process, as shown in Fig. 5(b) and Fig. 5(c), respectively. It is noticeable that the Brillouin intensity at the end of the sensing fiber is measured to be 0.66 mV while the root mean square of the noise floor was 0.046 mV, resulting in a high SNR of 11.6 dB. Using the exponential decay fitting on the measured gain trace, the maximal measurable sensing distance is estimated to be 66.9 km.
Fig. 5. Analysis on characteristics of spontaneous Brillouin scattering. (a) Measured Brillouin gain spectrum along the entire sensing fiber. (b) Distribution of Brillouin frequency shift. (c) Distributed Brillouin gain trace with an estimated maximal sensing distance using an exponential decay fitting.
The sensing performance is usually evaluated by the repeatability of successively measured Brillouin frequency referred to as frequency error, and the measurement repeatability or uncertainty is calculated by the standard deviation (1sigma) out of 5 consecutive measurements, as shown in Fig. 6. For this experiment, the adjustment of the NCO frequency scan and the acquisition of the Brillouin intensity trace is manually performed, which slowed down the measurement time to 23 minutes. So, any possible inhomogeneous temperature variation along the sensing fiber during this time may degrade the performance. In addition, the intrinsic Rayleigh back-scattering of the pulsed pump can be considered as another noise source since it is not electrically filtered out because the detection system is designed to analyze the heterodyne signal at DC with a 20 MHz bandwidth. Nevertheless, the repeatability of the sensing system is measured to be 4.5 MHz (following the exponential dependence of the frequency error over the distance [16], which is analogous to the definition in SEAFOM standard [15]) under the measurement conditions: Brillouin intensity trace averaging of 2000 and frequency scan step of 12.8 MHz, corresponding to a theoretical measurement time of ∼43 seconds. Using the figure-of-merit (FoM) formula established for BOTDA sensing system [16], our BOTDR system achieves the FoM of 3.0, confirming the good performance. Furthermore, according to the signal-to-noise ratio analysis formula [15], we can expect the repeatability of 1.6 MHz at 50 km position for 10 minutes measurement time.
Fig. 6. Measured Brillouin frequency uncertainty over the sensing fiber under measurement conditions of 2000 Brillouin intensity trace averaging and 12.8 MHz frequency scan step, resulting in the frequency error of 4.5 MHz at 50 km with 5 m spatial resolution.
To validate the spatial resolution of the sensing system, ∼5 m segment of the sensing fiber close to the end of the sensing fiber is thermally insulated at 47.5°C while the ambient room temperature is 22.7°C. Figure 7 illustrates the measured distribution of the Brillouin frequency shift over the last 50 m sensing fiber and the spectral profile of spontaneous Brillouin scattering at the position of 50.541 km at room temperature and at elevated temperature. As expected, an obvious increase in Brillouin frequency along ∼5-m-long fiber is clearly observed and localized at the heated part. The amount of Brillouin frequency upshift is measured to be 24 MHz, which has a good agreement to the temperature difference of 24.8°C, using the typical Brillouin temperature coefficient of 1 MHz/°C.
Fig. 7. Comparison of (a) the Brillouin frequency distribution and (b) the spontaneous Brillouin scattering spectrum with and without a hot spot at the sensing fiber end.
4. Conclusions
We have experimentally established a digital optical phase-locked loop, using an FPGA-based digital electronics platform, which was successfully applied to develop a simple and low-cost distributed sensing system with a high performance based on Brillouin scattering in optical fibers. A reliable phase-synchronization between two commercial semiconductor lasers is achieved with a locking bandwidth of 270 kHz, which is sufficient to realize a dynamic Brillouin sensing system, typically requiring a frequency scanning speed of <10 kHz. In addition, the frequency of the offset locking was largely tunable and flexible, simply by changing the NCO frequency. It allows to monitor the distributed temperature along the sensing fiber with a large range of Brillouin frequency variations in our experiment. Overall, we believe that such a versatile and cost-effective Brillouin sensing system will be of significant attractiveness for industrial applications.
Funding
Centre Suisse d'Electronique et de Microtechnique S.A..
Disclosures
The authors declare no conflicts of interest.
Data Availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. P. Lu, N. Lalam, M. Badar, B. Liu, B. T. Chorpening, M. P. Buric, and P. R. Ohodnicki, “Distributed optical fiber sensing: Review and perspective,” Appl. Phys. Rev. 6(4), 041302 (2019). [CrossRef]
2. M. F. Bado, J. R. Casas, A. Dey, and C. G. Berrocal, “Distributed Optical Fiber Sensing Bonding Techniques Performance for Embedment inside Reinforced Concrete Structures,” Sensors 20(20), 5788 (2020). [CrossRef]
3. A. Barrias, J. R. Casas, and S. Villalba, “A Review of Distributed Fiber Sensors for Civil Engineering Applications,” Sensors 16(5), 748 (2016). [CrossRef]
4. B. Wang, L. Wang, N. Guo, Z. Zhao, C. Yu, and C. Lu, “Deep neural networks assisted BOTDA for simultaneous temperature and strain measurement with enhanced accuracy,” Opt. Express 27(3), 2530 (2019). [CrossRef]
5. C. Hong, Y. Zhang, G. Li, M. Zhang, and Z. Liu, “Recent progress of using Brillouin distributed fiber optic sensors for geotechnical health monitoring,” Sens. Actuators, A 258, 131–145 (2017). [CrossRef]
6. T. Kurashima, T. Horiguchi, and M. Tateda, “Distribured-temperature sensing using stimulated Brillouin scattering in optical silica fibers,” Opt. Lett. 15(18), 1038–1040 (1990). [CrossRef]
7. M. Nikles, L. Thévenaz, and P. A. Robert, “Simple distribued fiber sensor based on Brillouin gain spectrum analysis,” Opt. Lett. 21(10), 758 (1996). [CrossRef]
8. S. Diaz, S. F. Mafang, M. Lopez-Amo, and L. Thévenaz, “A High-Performance Optical Time-Domain Brillouin Distributed Fiber Sensor,” IEEE Sens. J. 8(7), 1268–1272 (2008). [CrossRef]
9. L. Thévenaz, S. Le Floch, D. Alasia, and J. Troger, “Novel schemes for optical signal generation using laser injection locking with application to Brillouin sensing,” Meas. Sci. Technol. 15(8), 1519–1524 (2004). [CrossRef]
10. Y. Li, X. Bao, F. Ravet, and E. Ponomarev, “Distributed Brillouin sensor system based on offset locking of two distributed feedback lasers,” Appl. Opt. 47(2), 99–102 (2008). [CrossRef]
11. E. N. Ivanov, F. X. Esnault, and E. A. Donley, “Offset phase locking of noisy diode lasers aided by frequency division,” Rev. Sci. Instrum. 82(8), 083110 (2011). [CrossRef]
12. https://github.com/oscimp/oscimpDigital
13. T. Preuschoff, M. Schlosser, and G. Birkl, “Digital laser frequency and intensity stabilization based on the STEMlab platform (originally Red Pitaya),” Rev. Sci. Instrum. 91(8), 083001 (2020). [CrossRef]
14. K. Shimizu, T. Horiguchi, and T. Kurashim, “Coherent Self-Heterodyne Brillouin OTDR for Measurement of Brillouin Frequency Shift Distribution in Optical Fibers,” J. Lightwave Technol. 12(5), 730–736 (1994). [CrossRef]
15. SEAFOM-MSP-01, “Measurement Specification for Distributed Temperature Sensing,” (2016): https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved=2ahUKEwjE7Puml7PxAhXh_7sIHfpYAOgQFjAAegQIAxAD&url=https%3A%2F%2Fseafom.com%2F%3Fmdocs-file%3D80&usg=AOvVaw2fOTa9cJtkSV2LrZbN_VMp
16. M. Soto and L. Thévenaz, “Modeling and evaluating the performance of Brillouin distributed optical fiber sensors,” Opt. Express 21(25), 31347–31366 (2013). [CrossRef]
