Multiplexing fiber-optic Fabry&#x2013;Perot acoustic sensors using self-calibrating wavelength shifting interferometry

Qiang Liu; Zhenguo Jing; Zhenguo Jing; Yueying Liu; Ang Li; Zhenjie Xia; Wei Peng; Wei Peng

doi:10.1364/OE.381197

1. Introduction

Optical fiber acoustic sensors have been widely used in various fields, such as sound source localization [1], leakage detection [2,3], non-contact imaging [4], and photoacoustic spectroscopy [5], etc. Compared with conventional electrical acoustic sensors, they have the unique advantages of compact size, high sensitivity, anti-electromagnetic interference and remote sensing ability [6]. With the advancement of fiber-optic acoustic sensors and high-performance demodulation techniques, all-optical acoustic sensors have the potential to replace traditional electrical acoustic sensors in various areas, especially in the field of medical imaging requiring miniature sensor probes [7], or acoustic emission (AE) monitoring in harsh environments [8]. Diaphragm-based fiber-optic Fabry–Perot (FP) acoustic sensor is one of the most sensitive fiber-optic acoustic sensors reported to date [9]. The diaphragm vibrates with applied acoustic signal, resulting in length changes of the FP cavity [6]. Fiber-optic FP acoustic sensors have been applied in partial discharge detection [10], remote monitoring of gas micro-leakage [11] and non-contact ultrasonic detection in low-pressure carbon dioxide medium [12]. However, most of these applications are single-point acoustic detection. Since the quadrature points (Q-points) of multiple FP acoustic sensors are difficult to maintain consistent due to background temperature or pressure fluctuations, the most widely used Q-point intensity demodulation scheme [13,14] is invalid for multi-point acoustic detection. Phase interrogation schemes such as dual-wavelength or three-wavelength quadrature phase demodulation methods [12,15,16] have the advantages of large dynamic range and high-resolution, but they are also schemes for single-point acoustic interrogation. In practical applications, it is difficult for different FP sensors to maintain quadrature phase conditions at the same time. Fiber-optic FP sensors can be multiplexed in parallel assisted by cepstrum-division multiplexing (CDM) [17,18]. In that case, each FP sensor needs to have a different cavity length and the reflection spectrum is a combination of all FP interferometer contributions, which results in limited multiplexing channels and complex sensor fabrications. In addition, due to the limited acquisition speed of the spectrometer and the complex demodulation algorithm, the CDM method can only be used for relatively low frequency acoustic measurements. In 2018, Wang, et al. proposed a time division multiplexing (TDM) assisted coherent phase detection scheme, which is an optional phase demodulation technique for multiplexing diaphragm-based optical fiber acoustic sensors [19]. A 1*n coupler connects all the sensors in parallel with single mode fiber (SMF) delay line in different lengths for TDM. The optical sampling rate is about 60 kHz. However, the SMF delay line for each sensor probes is easily affected by irrelevant environmental vibrations, introducing phase demodulation errors. Flexible and high-resolution demodulation techniques for multiplexing fiber-optic FP acoustic sensors may greatly accelerate their industrial application, and, therefore, are highly desirable.

Phase-shifting interferometry (PSI) is a well-established high-precision, rapid and quantitative phase retrieval technique that has been widely used in surface profile testing, 3D shape measurement and live cell imaging [20–23]. It requires typically 3 to 8 phase-shifted frames for accurate digital phase demodulation, most commonly realized by a piezoelectric transducer (PZT). However, due to a series of problems such as hysteresis, nonlinearity and temperature drift, PZT may distort and cause phase shift errors. In addition, the mechanical phase shift frequency of PZT is limited, which hinders its application in high frequency acoustic sensing. Jiang et al. proposed a novel phase demodulation method for FP acoustic sensors based on phase shifting technique and birefringence crystals [24]. Four birefringence crystals with different thickness were used to obtain quadrature phase-shifted signals. Sinusoidal sonic signals of 21 kHz and 40 kHz can be successfully monitored by a four-step phase-shifting algorithm (PSA). Another potential solution is wavelength shifting interferometry (WSI), which generates the phase-shift by wavelength change of the light source. For a fixed wavelength-shift and an FP sensor with known initial cavity length, the corresponding phase-shift value can be accurately calculated based on the theoretical model of two-beam interference. For another FP sensor with a different cavity length, the phase-shift value can also be calculated, enabling simultaneous demodulation of multiple FP sensors. It should be noted that the cavity lengths of the multiple FP sensors are not required to be completely different or consistent, so that the manufacturing difficulty and the cost can be greatly reduced, which is essential for industrial applications. To date, WSI techniques are mainly applied in surface profile and thickness measurements [25,26], as well as relatively low-frequency quantitative phase imaging [27]. The development of fast-switching, high-quality tunable lasers makes WSI an attractive alternative of multiplexing fiber-optic FP acoustic sensors.

In this work, we propose a high-resolution cavity length demodulation technique for fiber-optic FP acoustic sensor array based on WSI. A monolithically integrated modulated grating Y-branch (MG-Y) laser was used for high-speed wavelength shifting. Cavity lengths of FP acoustic sensors can be simultaneously demodulated by a novel self-calibrating five-step PSA at a rate of 100 kHz. To the best of our knowledge, this is the first time to achieve multiplexing of fiber-optic FP acoustic sensors using WSI technique. Simultaneous acoustic detection of a four polyethylene terephthalate (PET) diaphragm based FP acoustic sensor array demonstrates its multiplexing performance. Moreover, we realize sound source location using the all-optical acoustic sensor array.

2. Principles

For WSI-based acoustic demodulation systems, the core component is the high-speed, stable and widely tunable light source. In this work, we utilize a MG-Y laser (ATLS 7500, AOC), a kind of vernier tuned distributed Bragg reflector (VT-DBR) laser, for high-speed wavelength switching to introduce phase shifts in the wavelength domain. It has high wavelength stability and repeatability due to precise electronic tuning without mechanical movement [15,28]. The output wavelength of the MG-Y laser is determined by the injection currents of the right reflector section, the left reflector section and the phase section, and the output intensities are adjusted by fine-tuning the injection current of the semiconductor optical amplifier (SOA) section [15]. Herein, we realize high precision wavelength control in the range of 1527 ∼ 1567 nm utilizing a field-programmable gate array (FPGA), and adjust the output intensity at each wavelength to be consistent (10.5 dBm) [15]. Considering that the output intensity at each wavelength has been adjusted in real time, there is no need to normalize the interference fringe or subtract the direct current (DC) term in WSI, which greatly reduces the complexity of demodulation and improves the robustness of the system.

In standard PSI, a sequence of N phase-shifting interferograms with constant phase shifts are required for phase retrieval. Usually a minimum of three phase-shifting interferograms are needed. To date, numerous PSAs have been proposed or extended [20,29], such as 3-step, 4-step, 5-step and least-square algorithm, etc. In general, more phase shift steps result in higher phase extraction accuracy and larger phase-shift error tolerance. However, it also comes with slower demodulation speed and increased complexity. In order to strike a balance between the demodulation speed and the phase retrieval accuracy, we used a modified 5-step Hariharan PSA [30] in this work, which is one of the most widely used PSAs with high phase retrieval accuracy. Further, the phase-shift step θ is calibrated in real time to enhance its robustness for applications in harsh environments.

2.1 Self-calibrating 5-step PSA for single-point demodulation

The WSI system for the interrogation of a single fiber-optic FP acoustic sensor is shown in Fig. 1. Figure 1(a) illustrates the schematic diagram. A MG-Y tunable laser is used as the light source. Synchronized wavelength switching and data acquisition are implemented by an FPGA. The inset shows the diagram of the FP acoustic sensor. Polyethylene terephthalate (PET) diaphragms with thickness of 6 µm are used to fabricate the FP acoustic sensors. A low-finesse FP cavity is formed between the cleaved end face of the fiber and the inner surface of the PET diaphragm. We used capillaries to support and align the fiber. The capillary used for PET diaphragm attachment has an inner diameter of 3.4 mm. As shown in Fig. 1(b), the five operating wavelengths are sequentially switched. The intensities corresponding to each wavelengths are uploaded to a computer for real-time phase retrieval. Figure 1(c) presents the flow chart of the demodulation process, the principles and details will be described below.

Fig. 1. Single-point acoustic demodulation system based on wavelength shifting interferometry (WSI). (a) Schematic diagram of the system. (Inset) Diagram of the FP acoustic sensor with a polyethylene terephthalate (PET) diaphragm. (b) Five-step wavelength modulation diagram. (c) Single-point demodulation flowchart.

Download Full Size | PDF

The FP acoustic sensor used here can be regarded as a low-finesse two-beam interferometer. Intensity of the reflected interferential light at a specific wavelength λ can be expressed as

(1)$$I(\lambda )= A + B\cos [{\varphi (\lambda )+ {\varphi_0}} ],$$

where A is the DC component of the interferometric fringe, B is the fringe visibility, φ₀ is the initial phase, which is seen as a constant. The phase corresponding to λ, φ(λ), can be expressed as

(2)$$\varphi (\lambda )= \frac{{4n\pi }}{\lambda }{L_0},$$

where n is the refractive index of the EFPI cavity (here n = 1), L₀ is the initial cavity length. So that the phase-shift θ between two adjacent operating wavelengths can be written as

(3)$$\theta = {\varphi _{i + 1}} - {\varphi _i} = 4n\pi {L_0}\left( {\frac{1}{{{\lambda_{i + 1}}}} - \frac{1}{{{\lambda_i}}}} \right) \approx 4n\pi {L_0}\frac{{\Delta \lambda }}{{{\lambda _3}^2}},({i = 1,2,3,4} )$$

where i denotes the sequence number of wavelength shifting, λ₃ is the center wavelength of the five operating wavelengths. Based on Eq. (3), target phase-shift θ can be introduced by equivalent wavelength-shift Δλ, which can be expressed as

(4)$$\Delta \lambda = \frac{{{\lambda _3}^2\theta }}{{4n\pi {L_0}}}.$$

For convenience, we set λ₃ at the center wavelength of the tunable range, 1547 nm, the five operating wavelengths for WSI can be determined based on θ and L₀. Herein, for single-point demodulation, the preset phase-shift step θ is determined as π/2 radians to achieve the best sensitivity and make it insensitive to phase-shift errors [31]. Equation (4) can be converted to

(5)$$\Delta \lambda = \frac{{{\lambda _3}^2}}{{8n{L_0}}}.$$

The interference spectrum can be easily obtained by linear wavelength scanning of the MG-Y laser (covering from 1527 nm to 1567 nm). The initial cavity length L₀ is measured by a cross correlation algorithm [32]. Based on Eqs. (1)–(3), the intensities at the five operating wavelengths can be expressed as

(6)$$\left\{ \begin{array}{l} {I_1} = A + B\cos [{({{\varphi_3} + {\varphi_0}} )- 2\theta } ]\\ {I_2} = A + B\cos [{({{\varphi_3} + {\varphi_0}} )- \theta } ]\\ {I_3} = A + B\cos ({{\varphi_3} + {\varphi_0}} )\\ {I_4} = A + B\cos [{({{\varphi_3} + {\varphi_0}} )+ \theta } ]\\ {I_5} = A + B\cos [{({{\varphi_3} + {\varphi_0}} )- 2\theta } ]\end{array} \right.$$

Consequently, the time-varying phase value corresponding to λ₃ can be calculated based on a Hariharan 5-step PSA,

(7)$$\Phi = {\varphi _3} + {\varphi _0} = \arctan \left[ {\frac{{1 - \cos 2\theta }}{{\sin \theta }} \cdot \left( {\frac{{{I_2} - {I_4}}}{{2{I_3} - {I_1} - {I_5}}}} \right)} \right].$$

The problem of phase winding will be encountered because of the arctangent operation, phase jump occurs when the original Φ approaches ±π/2. Based on the continuity of phase change during acoustic detection, real-time phase compensation can be achieved. According to Eqs. (2) and (7), the cavity length change ΔL can be calculated as

(8)$$\Delta L = \frac{{{\lambda _3} \cdot \Delta {\varphi _3}}}{{4n\pi }} = \frac{{{\lambda _3} \cdot \Delta \Phi }}{{4n\pi }}.$$

The real-time cavity length of the FP acoustic sensor, L_t, can be expressed as

(9)$${L_t} = {L_0} + \Delta L.$$

In the time-varying waveform of L_t, a DC component, L_DC, and an AC component, L_AC, are included, where L_AC corresponds to the applied acoustic signal, and the low-frequency variation of L_DC demonstrates ambient temperature or pressure fluctuations. Large-amplitude changes in L_DC may result in phase-shift deviation from the preset phase-shift step (θ =π/2), leading to phase retrieval errors. It is one of the main concerns to be addressed for applications in harsh environments. In this paper, creatively, we use L_DC for the real-time calibration of the phase-shift step. According to Eq. (3), the calibrated phase-shift step θ can be written as

(10)$$\theta \approx 4n\pi {L_{DC}}\frac{{\Delta \lambda }}{{{\lambda _3}^2}}.$$

Briefly, for single sensor demodulation, linear wavelength scanning is first performed to obtain the initial cavity length L₀ and determine the five operating wavelengths for generating five quadrature phase-shift (θ =π/2). In five-step WSI, corresponding light intensities are synchronously acquired to calculate the phase change ΔΦ and the real-time cavity length L_t. Five selected wavelengths are sequentially switched at a rate of 500 kHz, resulting in a cavity length sampling frequency of 100 kHz. Further, the DC term L_DC is used for real-time calibration of phase-shift step, making the system robust in applications involving large environmental perturbations.

2.2 Simultaneous demodulation of fiber-optic FP acoustic sensor array

If we know the phase-shift step corresponding to each sensing channel, the 5-step WSI technique using a single MG-Y laser can be easily extended for multiplexing fiber-optic FP acoustic sensors. Fortunately, the phase-shift step can be calculated by Eq. (3) for a known cavity length and a fixed wavelength-shift. The schematic of the WSI-based multiplexing system is illustrated in Fig. 2. A 1*N coupler is used to achieve spatial-division-multiplexing (SDM) of multiple sensors. The signals reflected from each sensor are detected by multiple PDs and then uploaded to a computer for data processing. Herein, we use the average cavity length L_avg for the determination of the five operating wavelengths. Equation (5) for generating π/2 phase-shift can be converted to

(11)$$\Delta \lambda = \frac{{{\lambda _3}^2}}{{8n{L_{avg}}}}$$

The phase-shift step of each channel can be calculated as

(12)$${\theta _m} \approx 4n\pi {L_m}\frac{{\Delta \lambda }}{{{\lambda _3}^2}} = \frac{{{L_m}}}{{{L_{avg}}}} \cdot \frac{\pi }{2},({m = 1,2,3,\ldots ., n} )$$

where m represents the channel sequence number, L_m is the corresponding initial cavity length. In theory, the phase-shift step can take an arbitrary value if we know it in advance [31]. However, in this work, in order to maintain good sensitivity and stability of all sensing channels, the phase-shift steps are controlled within the range of [π/3, 2π/3]. Therefore, the acceptable cavity length difference is approximately ± 1/3 L_avg. This is a relatively large tolerance range, which greatly reduces the manufacturing difficulty. As single-point demodulation, the cavity length variations of each sensors can be demodulated based on Eqs. (7)–(9), and the phase-shift step θ_m is also calibrated in real-time based on Eq. (10).

Fig. 2. Schematic diagram of the WSI based multi-point acoustic demodulation system.

Download Full Size | PDF

Figure 3 exhibits the flowchart of the self-calibrating 5-step WSI for simultaneous demodulation of an N-channel sensor array. As the single-sensor demodulation process, full-spectrum scanning is required to obtain the initial cavity length of each sensor at the beginning. Next, five operating wavelengths are determined based on L_avg. Phase-shift step of each channel is calculated based on the corresponding initial cavity length and is calibrated in real-time based on the L_DC. The proposed WSI technique can realize simultaneous cavity length demodulation of large-scale fiber-optic acoustic sensor array with high precision and large dynamic range.

Fig. 3. Flowchart of the self-calibrating 5-step WSI for multi-point demodulation.

Download Full Size | PDF

3. Experimental results and discussion

We established a four-channel acoustic sensing system to demonstrate the validity and performance of the proposed WSI demodulation scheme. Four sensor probes are put in an anechoic test box for simultaneous acoustic sensing. A speaker connected to the signal generator is used to generate acoustic signals of different frequencies. Linear wavelength scanning from 1567 nm to 1527 nm at an interval of 8 pm was first performed to obtain the initial cavity length of each channel. Figure 4 shows the interference spectrums of the four sensors. It can be seen that the fringe visibility and DC component of the interference spectrums are different owing to the influences of different insertion loss of couplers, different gains and responsivities of the PDs and different sensor cavity lengths, etc. Since the PD used in channel-3 (Thorlabs, USA) is inconsistent with the PDs (Keyang Photonics, China) used in the other three channels, the amplitudes of Channel-3 is quite different from the other three channels. It will not influence the effectiveness of the proposed WSI demodulation methods. Based on a white light interference absolute cavity length demodulation method [32,33], the initial cavity lengths of the four sensors are calculated as 152.845 µm, 147.205 µm, 126.481 µm and 136.766 µm, respectively. The average cavity length is 140.824 µm. In order to introduce five-step π/2 phase-shifts corresponding to the average cavity length, the five operating wavelengths are determined as 1551.256 nm, 1549.128 nm, 1547.000 nm, 1544.872 nm and 1542.744 nm. Corresponding actual initial phase-shift steps for the four channels are calculated as 0.542π, 0.523π, 0.449π and 0.485π (97.7°, 94.1°, 80.8°, and 87.4°). The phase-shift steps are all in the range of [π/3, 2π/3], ensuring the demodulation accuracy.

Fig. 4. Interference spectrums of the four FP acoustic sensors obtained by linear wavelength scanning. Their cavity lengths are 152.845 µm, 147.205 µm, 126.481 µm and 136.766 µm, respectively.

Download Full Size | PDF

Subsequently, five operating wavelengths are sequentially switched. Corresponding light intensities at different wavelengths are detected and then separated in the time domain for cavity length demodulation. All the four channels are demodulated simultaneously. In particular, demodulation results of Sensor-1 are extracted to demonstrate the system performance. Figure 5 displays the time domain responses and demodulation results of Sensor-1. When no acoustic signal is applied, the intensity waveforms corresponding to λ₁ ∼ λ₅ are shown in Fig. 5(a). Cavity length variations are then calculated based on the self-calibrating 5-step PSA. The standard deviation of demodulated cavity length variations is about 0.3 nm, demonstrating the stability of demodulation results. Figures 5(b) and 5(c) display the original intensity waveforms and demodulated cavity length variations at 4 kHz, 1.0 Pa acoustic signals. Corresponding power spectrum is presented in Fig. 5(d) with a signal to noise ratio (SNR) of about 70.2 dB. It can be found that the demodulation result is consistent with the applied acoustic signal. The minimum detectable pressure (MDP) is calculated to be 68.9 µPa/Hz^1/2.

Fig. 5. Time domain responses and demodulation results of Sensor-1: (a) Extracted original intensity signals corresponding to λ₁ ∼ λ₅ with no applied acoustic signals. (b) Extracted original intensity signals at 4 kHz applied acoustic signals. (c) Demodulated cavity length variations. (d) Power spectrum in the frequency domain.

Download Full Size | PDF

Then the FP acoustic sensors were used to detect 8 kHz and 12 kHz acoustic signals. The corresponding time domain cavity length variations and power spectrums are shown in Figs. 6(a) and 6(b). Experimental results prove its ability to demodulate acoustic signals of different frequencies. Figure 6(c) plots the frequency response of Sensor-1. Limited by the performance of the loud speaker, the tested frequency range is 1 kHz to 16 kHz. Theoretically, considering the demodulation frequency of 100 kHz and the Nyquist sampling theorem, the maximum frequency that can be demodulated is approximately 50 kHz.

Fig. 6. Demodulation results at 8 kHz and 12 kHz acoustic signals and the frequency response of Sensor-1. (a) Cavity length waveforms and power spectrum at 8 kHz. (b) Cavity length waveforms and power spectrum at 12 kHz. (c) Frequency response of Sensor-1.

Download Full Size | PDF

Taking advantages of the cavity length demodulation scheme, a large dynamic range of acoustic pressure can be detected. Figure 7(a) shows the acoustic pressure response of the Sensor-1 at 4 kHz. Larger acoustic pressure results in a larger cavity length amplitude. The relationships between the amplitude of cavity length variations and the applied acoustic pressure have been fitted and plotted in Fig. 7(b). The applied sound pressure varies between 0 Pa and 2.86 Pa by adjusting the output voltage of the signal generator. It can be found that the responses of all the four channels have a good linearity (R² ≥ 0.999 for all channels) and similar sensitivity. The sensitivity of Sensor-1 is calculated to be 102.8 nm/Pa, and the sensitivity of the other three sensors is calculated as 134.4 nm/Pa, 108.7 nm/Pa and 122.9 nm/Pa, respectively. The minor differences in sensitivity may be due to differences in the adhesion of diaphragms [19], the position or orientation of the sensor, and so on.

Fig. 7. Acoustic pressure responses of the FP acoustic sensor at 4 kHz: (a) Demodulated cavity length variations of Sensor-1 at different acoustic pressures. (b) Fitting curves of acoustic pressure responses of the four sensors.

Download Full Size | PDF

Sound source localization is a typical application for multipoint acoustic sensing. All-optical acoustic sensor array exhibits significant superiority for sound source localization in harsh environments, and has attracted much research interests in recent years [1,34]. Fiber-optic FP acoustic sensors are suitable for sound source localization for their probe-type structure and high sensitivity. We used the WSI-based four channel acoustic demodulation system to realize two-dimensional (2D) sound source localization. As shown in Fig. 8, the four fiber-optic sensor probes are placed at four points on the XY coordinate axis. Each sensor is 45 cm away from the origin point O. Coordinates of the four sensors are Sensor-1 (45, 0), Sensor-2 (0, 45), Sensor-3 (−45, 0) and Sensor-4 (0, −45), respectively. A hyperbolic localization algorithm is utilized to locate the sound source [35]. The time differences of arrival (TDOA) measured by two pairs of sensors along the orthogonal axes are used to determine two hyperbolic curves. For Sensor-1 and Sensor-3 on the horizontal axis, the sound source location (x, y) should satisfy:

(13)$$\frac{{{x^2}}}{{{a_h}^2}} - \frac{{{y^2}}}{{{b_h}^2}} = 1$$

Where

(14)$$\left\{ \begin{array}{l} {a_h} = {v_s} \cdot \frac{{\Delta {t_{13}}}}{2}\\ {b_h}^2 = {d^2} - {a_h}^2\\ \Delta {t_{13}} = {t_3} - {t_1} \end{array} \right.$$

v_s is the sound velocity in the air, which is 340 m/s. t₁ and t₃ represent the time at which the acoustic waves reach the sensor. Δt₁₃ is the TDOA between Sensor-1 and Sensor-3. d is the distance between the sensor and the origin, which is 45 cm in this experiment. A hyperbolic curve can be determined based on Sensor-1 and Sensor-3. Similarly, another hyperbolic curve can be determined based on Sensor-2 and Sensor-4 on the vertical axis. The intersection of the two hyperbolic curves is calculated as the coordinates of the sound source location.

Fig. 8. Layout of the four fiber-optic FP acoustic sensors and the positioning principle based on a hyperbolic localization algorithm. Each sensor is 45 cm away from the origin point O. The coordinates of the four sensors are Sensor-1 (45, 0), Sensor-2 (0, 45), Sensor-3 (−45, 0) and Sensor-4 (0, −45), respectively. Red curves are determined by the time differences of arrival (TDOA) between Sensor-1 and Sensor-3, and blue curves are determined by the TDOA between Sensor-2 and Sensor-4. Their intersection is the estimated sound source position.

Download Full Size | PDF

Figure 9 plots the time-domain waveforms detected by the four-sensor array when the actual sound source position is (30, 15). Herein, three times the noise value is set as the threshold, and the arrival time is determined when the first peak or valley occurs. The TDOA between Sensor-1 and Sensor-3 was calculated as 1.66 ms, and that for Sensor-2 and Sensor-4 was 0.68 ms. The estimated coordinate of the sound source position is (30.42, 14.11), which is in good agreement with the actual position (30, 15).

Fig. 9. Typical time-domain waveforms detected by the four-sensor array when the actual sound source position is (30, 15). The TDOA between Sensor-1 and Sensor-3 was calculated as 1.66 ms and that for Sensor-2 and Sensor-4 was 0.68 ms.

Download Full Size | PDF

Considering of the 100 kHz optical sampling frequency, the theoretical spatial resolution of the sound source localization is about 0.34 cm. In order to evaluate the positioning accuracy of the proposed sound source localization system, we conducted localization tests at 25 different positions. Localization results of 6 repeated measurements are shown in Fig. 10. The estimated sound source positions are marked as red crosses, and the actual sound source positions are presented as black circles. Positioning errors of all measurements are no larger than 2.42 cm, demonstrating its accuracy. The positioning results show that effective 2D sound source localizations can be realized based on the all-optical four-sensor system.

Fig. 10. Positioning results of 25 different positions.

Download Full Size | PDF

4. Conclusion

In this paper, we propose a WSI-based cavity length demodulation technique and demonstrate its application in multiplexing fiber-optic FP acoustic sensors. A widely tunable monolithic MG-Y laser is used for high-speed wavelength shifting, introducing known phase-shifts for large-scale acoustic sensor array. To maintain high sensitivity and good stability in all the sensing channels, the phase-shift steps are controlled in the range of [π/3, 2π/3], which corresponds to an acceptable cavity length difference of approximately ± 1/3 L_avg. It greatly reduces the difficulty of sensor fabrication and mass production. Based on a self-calibrating 5-step PSA, a 100 kHz cavity length sampling frequency is obtained. We experimentally demonstrated an all-optical four-sensor acoustic demodulation system based on this WSI technique, and used it for 2D sound source localization. Experimental results show that the positioning accuracy is about 2.42 cm. Considering its all-optical structure, compact size, high precision and good stability, this WSI-based multi-point acoustic detection system has great potentials in industrial applications.

Funding

National Natural Science Foundation of China (61520106013, 61727816); Fundamental Research Funds for the Central Universities (DUT18ZD215).

Disclosures

The authors declare no conflicts of interest.

References

1. J. Wang, H. Peng, P. Zhou, J. Guo, B. Jia, and H. Wu, “Sound source localization based on Michelson fiber optic interferometer array,” Opt. Fiber Technol. 51, 112–117 (2019). [CrossRef]

2. S. Masuzawa, S. Okazaki, Y. Maru, and T. Mizutani, “Catalyst-type-an optical fiber sensor for hydrogen leakage based on fiber Bragg gratings,” Sens. Actuators, B 217, 151–157 (2015). [CrossRef]

3. L. Ren, T. Jiang, Z. G. Jia, D. S. Li, C. L. Yuan, and H. N. Li, “Pipeline corrosion and leakage monitoring based on the distributed optical fiber sensing technology,” Measurement 122, 57–65 (2018). [CrossRef]

4. Q. Rong, R. Zhou, Y. Hao, X. Yin, Z. Shao, T. Gang, and X. Qiao, “Ultrasonic sensitivity-improved Fabry–Perot interferometer using acoustic focusing and its application for noncontact imaging,” IEEE Photonics J. 9(3), 1–11 (2017). [CrossRef]

5. K. Chen, S. Liu, B. Zhang, Z. Gong, Y. Chen, M. Zhang, H. Deng, M. Guo, F. Ma, F. Zhu, and Q. Yu, “Highly sensitive photoacoustic multi-gas analyzer combined with mid-infrared broadband source and near-infrared laser,” Opt. Lasers Eng. 124, 105844 (2020). [CrossRef]

6. H. Liao, P. Lu, L. Liu, S. Wang, W. Ni, X. Fu, D. Liu, and J. Zhang, “Phase Demodulation of Short-Cavity Fabry–Perot Interferometric Acoustic Sensors With Two Wavelengths,” IEEE Photonics J. 9(2), 1–9 (2017). [CrossRef]

7. A. Rosenthal, S. Kellnberger, D. Bozhko, A. Chekkoury, M. Omar, D. Razansky, and V. Ntziachristos, “Sensitive interferometric detection of ultrasound for minimally invasive clinical imaging applications,” Laser Photonics Rev. 8(3), 450–457 (2014). [CrossRef]

8. P. Wei, X. Han, and D. Xia, “Measurement Method for the Velocity of Acoustic Emission Wave in Liquid Nitrogen,” IEEE Trans. Ind. Electron. 65(10), 8232–8238 (2018). [CrossRef]

9. L. Liu, P. Lu, H. Liao, S. Wang, W. Yang, D. Liu, and J. Zhang, “Fiber-Optic Michelson Interferometric Acoustic Sensor Based on a PP/PET Diaphragm,” IEEE Sens. J. 16(9), 3054–3058 (2016). [CrossRef]

10. W. Si, C. Fu, D. Li, H. Li, P. Yuan, and Y. Yu, “Directional Sensitivity of a MEMS-Based Fiber-Optic Extrinsic Fabry–Perot Ultrasonic Sensor for Partial Discharge Detection,” Sensors 18(6), 1975 (2018). [CrossRef]

11. K. Chen, M. Guo, S. Liu, B. Zhang, H. Deng, Y. Zheng, Y. Chen, C. Luo, L. Tao, M. Lou, and Q. Yu, “Fiber-optic photoacoustic sensor for remote monitoring of gas micro-leakage,” Opt. Express 27(4), 4648–4659 (2019). [CrossRef]

12. J. Jiang, T. Zhang, S. Wang, K. Liu, C. Li, Z. Zhao, and T. Liu, “Noncontact Ultrasonic Detection in Low-Pressure Carbon Dioxide Medium Using High Sensitivity Fiber-Optic Fabry–Perot Sensor System,” J. Lightwave Technol. 35(23), 5079–5085 (2017). [CrossRef]

13. J. Ma, M. Zhao, X. Huang, H. Bae, Y. Chen, and M. Yu, “Low cost, high performance white-light fiber-optic hydrophone system with a trackable working point,” Opt. Express 24(17), 19008–19019 (2016). [CrossRef]

14. X. Mao, S. Yuan, P. Zheng, and X. Wang, “Stabilized Fiber-Optic Fabry–Perot Acoustic Sensor Based on Improved Wavelength Tuning Technique,” J. Lightwave Technol. 35(11), 2311–2314 (2017). [CrossRef]

15. Q. Liu, Z. Jing, A. Li, Y. Liu, Z. Huang, Y. Zhang, and W. Peng, “Common-path dual-wavelength quadrature phase demodulation of EFPI sensors using a broadly tunable MG-Y laser,” Opt. Express 27(20), 27873–27881 (2019). [CrossRef]

16. Q. Liu, Z. Jing, Y. Liu, A. Li, Y. Zhang, Z. Huang, M. Han, and W. Peng, “Quadrature phase-stabilized three-wavelength interrogation of a fiber-optic Fabry–Perot acoustic sensor,” Opt. Lett. 44(22), 5402–5405 (2019). [CrossRef]

17. D. Tosi, “Simultaneous Detection of Multiple Fiber-Optic Fabry–Perot Interferometry Sensors With Cepstrum-Division Multiplexing,” J. Lightwave Technol. 34(15), 3622–3627 (2016). [CrossRef]

18. Z. Yu and A. Wang, “Fast Demodulation Algorithm for Multiplexed Low-Finesse Fabry–Pérot Interferometers,” J. Lightwave Technol. 34(3), 1015–1019 (2016). [CrossRef]

19. J. Wang, F. Ai, Q. Sun, T. Liu, H. Li, Z. Yan, and D. Liu, “Diaphragm-based optical fiber sensor array for multipoint acoustic detection,” Opt. Express 26(19), 25293–25304 (2018). [CrossRef]

20. C. Zuo, S. Feng, L. Huang, T. Tao, W. Yin, and Q. Chen, “Phase shifting algorithms for fringe projection profilometry: A review,” Opt. Lasers Eng. 109, 23–59 (2018). [CrossRef]

21. Y. Park, C. Depeursinge, and G. Popescu, “Quantitative phase imaging in biomedicine,” Nat. Photonics 12(10), 578–589 (2018). [CrossRef]

22. X. Wu, H. Wei, H. Zhang, L. Ren, Y. Li, and J. Zhang, “Absolute distance measurement using frequency-sweeping heterodyne interferometer calibrated by an optical frequency comb,” Appl. Opt. 52(10), 2042–2048 (2013). [CrossRef]

23. X. Wu, J. Zhang, H. Wei, and Y. Li, “Phase-shifting interferometer using a frequency-tunable diode laser calibrated by an optical frequency comb,” Rev. Sci. Instrum. 83(7), 073107 (2012). [CrossRef]

24. J. Jiang, Z. Zhao, S. Wang, K. Liu, Y. Huang, C. Shan, H. Xiao, and T. Liu, “Optical fiber Fabry-Perot interferometer based on phase-shifting technique and birefringence crystals,” Opt. Express 26(17), 21606–21614 (2018). [CrossRef]

25. Y. Kim, N. Sugita, and M. Mitsuishi, “Measurement of surface profile and thickness of multilayer wafer using wavelength-tuning fringe analysis,” Precis. Eng. 52, 130–137 (2018). [CrossRef]

26. Y. Sun, H. Shen, X. Li, J. Li, J. Gao, and R. Zhu, “Wavelength-tuning point diffraction interferometer resisting inconsistent light intensity and environmental vibration: application to high-precision measurement of a large-aperture spherical surface,” Appl. Opt. 58(5), 1253–1260 (2019). [CrossRef]

27. S. Chen, C. Li, and Y. Zhu, “Sensitivity evaluation of quantitative phase imaging: a study of wavelength shifting interferometry,” Opt. Lett. 42(6), 1088–1091 (2017). [CrossRef]

28. J. Park, E. F. Carbajal, X. Chen, J. S. Oghalai, and B. E. Applegate, “Phase-sensitive optical coherence tomography using an Vernier-tuned distributed Bragg reflector swept laser in the mouse middle ear,” Opt. Lett. 39(21), 6233–6236 (2014). [CrossRef]

29. J. Li, L. Zhong, S. Liu, Y. Zhou, J. Xu, J. Tian, and X. Lu, “An advanced phase retrieval algorithm in N-step phase-shifting interferometry with unknown phase shifts,” Sci. Rep. 7(1), 44307 (2017). [CrossRef]

30. P. Hariharan, B. F. Oreb, and T. Eiju, “Digital phase-shifting interferometry: a simple error-compensating phase calculation algorithm,” Appl. Opt. 26(13), 2504–2506 (1987). [CrossRef]

31. S. Chen and Y. Zhu, “Phase sensitivity evaluation and its application to phase shifting interferometry,” Methods 136, 50–59 (2018). [CrossRef]

32. J. Zhenguo and Y. Qingxu, “White light optical fiber EFPI sensor based on cross-correlation signal processing method,” in Proc. 6th Int. Symp. Test and Measurement, 2005), 3509.

33. X. Zhou and Q. Yu, “Wide-Range Displacement Sensor Based on Fiber-Optic Fabry–Perot Interferometer for Subnanometer Measurement,” IEEE Sens. J. 11(7), 1602–1606 (2011). [CrossRef]

34. B.-W. Jang and C.-G. Kim, “Impact localization of composite stiffened panel with triangulation method using normalized magnitudes of fiber optic sensor signals,” Compos. Struct. 211, 522–529 (2019). [CrossRef]

35. T. Fu, Y. Liu, K.-t. Lau, and J. Leng, “Impact source identification in a carbon fiber reinforced polymer plate by using embedded iber optic acoustic emission sensors,” Composites, Part B 66, 420–429 (2014). [CrossRef]

Multiplexing fiber-optic Fabry–Perot acoustic sensors using self-calibrating wavelength shifting interferometry

Abstract

1. Introduction

2. Principles

2.1 Self-calibrating 5-step PSA for single-point demodulation

2.2 Simultaneous demodulation of fiber-optic FP acoustic sensor array

3. Experimental results and discussion

4. Conclusion

Funding

Disclosures

References

Cited By

Figures (10)

Equations (14)

Optics Express