Open Access
Add to BibSonomy
Abstract
A novel fiber optic extrinsic Fabry-Perot interferometric (EFPI) ultrasonic sensor with two resonant frequencies for detecting the partial discharges (PDs) in switchgear is demonstrated. The key sensing element consists of two 5-µm-thickness and beam-supported silicon diaphragms, whose natural frequencies are designed differently to enable the sensor to achieve the resonant responses at two different frequencies, thus obtaining a broadened frequency response. The sensing element is fabricated by employing the microelectromechanical systems (MEMS) technology on a silicon-on-insulator (SOI) wafer. The experimental results show that the sensor possesses two resonant frequencies of 31 kHz and 63 kHz, and obviously, shows a highly sensitive frequency response over a broader range compared with the approach composed of a single sensing diaphragm with only one resonant frequency. The noise-limited minimum detectable ultrasonic pressure (MDUP) reaches 251 µPa/Hz1/2@ 31 kHz and 316 µPa/Hz1/2@ 63 kHz, respectively.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Switchgear works as a main equipment of electric power industry, and serves under relatively harsh electromagnetic environments accompanied by mechanical vibration, as well as other chemical and physical interactions, which tend to cause an insulation damage. And it is one of the greatest security risks, potentially leading to huge economic loss and casualties. Fortunately, this can be detected in advance by measuring the partial discharges (PDs) thus generated at the same time [1–3].
In practice, there are three main methods serving to detect PDs in switchgear, which are transient earth voltage (TEV), ultrahigh frequency (UHF) and ultrasonic methods, respectively [4–5]. TEV with high sensitivity and reliability is a powerful technique for noninvasively checking the condition of switchgears. However, the high requirement set for the operational environment due to the low resistance against the electromagnetic interference (EMI) greatly limits its practical application [6]. UHF seems to be a suitable method for continuously on-line monitoring, but it is difficult to ascertain the location of PDs and requires a high cost [7]. Ultrasonic method works as an extensively-applied PDs detecting method, one obvious advantage of which is that the position of a PD can be located by the phase delay or the amplitude attenuation of the ultrasonic waves. In addition, ultrasonic method shows the great advantage for noise immunity during the on-line PDs detection [8–9].
The fiber optic extrinsic Fabry-Perot interferometer (EFPI) sensor has become an ideal candidate for detecting the ultrasonic waves generated by PDs, since it exhibits many inherent merits, such as high sensitivity, low cost, small size, light weight, high frequency response, electrical insulation, and immunity to EMI noise [10–14]. The ultrasonic wave is a dynamic signal that needs to be transformed into a detectable physical parameter through the sensing diaphragm of the EFPI sensor, making it a key important component determining the performance of such sensor. At present, the sensing diagram still generally adopts circle diaphragm, and the method for improving its sensitivity mainly focuses on how to reduce its thickness. For example, graphene, silver, silicon and other materials were utilized to achieve low-thickness diaphragms, ranging from 100 nm to 450 nm [15–17]. However, when the diaphragm thickness is greatly reduced to achieve the required performance, the sensor is difficult to detect the ultrasonic waves due to low natural frequency of sensing diaphragms. Qichao Chen et al. proposed a series of EFPI ultrasonic sensors using circular silica diaphragm for PDs detection, and the diaphragm thickness was at least 20 µm [18]. Similarly, the diaphragm thickness in most reported research for PDs detection using the EFPI sensor is large, and usually, the intact circular structure is adopted [4–5].
In this paper, a fiber optic EFPI ultrasonic sensor with two resonant frequencies for PDs detection in switchgear is proposed. To reduce the damping of mechanical vibration and improve the resonant sensitivity, the sensing diaphragm with a thickness of 5 µm is supported by four beams. Two such beam-supported diaphragms with different natural frequencies are designed and fabricated on the sensing element of EFPI sensor, which can enable the sensor to achieve two different resonant responses, thus obtaining a broadened frequency response to improve the detecting capability for weak PDs. Finally, the sensing element is successfully fabricated on a silicon-on-insulator (SOI) wafer by using the Microelectromechanical Systems (MEMS) technology. The testing result shows that the noise-limited minimum detectable ultrasonic pressure (MDUP) of the developed sensor reaches 251 µPa/Hz1/2 and 316 µPa/Hz1/2 at the two resonant frequencies, respectively. The damping ratio of diaphragm vibration is around 0.1, and the sensitivity at resonant frequency increases at least 16 dB compared with that at twice or half the frequency. To the best of our knowledge, the EFPI sensor using two sensing diaphragms with different resonant frequencies for PDs detection has not been reported. The developed sensor has the great potential to enhance the capability for detecting the initially weak PDs.
2. Design and fabrication
2.1 Structural design
As the circular sensing diaphragm is excited by ultrasonic waves, its sensitivity sd can be expressed as:
According to Eq. (1), when the natural frequency fn of diaphragm is equal to the frequency f of ultrasonic wave and the damping ratio ζ is very small, the sensitivity of the diaphragm will be magnified infinitely, as shown in Fig. 1, which presents the sensitivity of circular diaphragm with different damping ratios (A is assumed to be 1 since it is only related to the materials and the geometrical parameters of the diaphragm). For the sake of decreasing the damping ratio and improving the resonant sensitivity of the sensor, the sensing diaphragm in this research adopts the four-beam supported structure as shown in Fig. 2. Another advantage of the beam-supported diaphragm is that it possesses a larger A compared with the intact circular diaphragm in our previous research [19]. In a word, the beam-supported diaphragm with a larger A and a smaller ζ is very promising to reveal a high sensitivity at the resonant frequency, thus resulting in a low MDUP which will improve the detection capability of weak PDs. In general, the characteristic frequencies of ultrasonic waves generated by PDs caused by different insulation damages are unknown. Therefore, in this paper, two beam-supported silicon diaphragms with the natural frequency of 30 kHz and 60 kHz are designed and fabricated on the sensing element to achieve the resonant responses at two different frequencies, thus obtaining a broadened frequency response to improve the capability for detecting the initially weak PDs. Finite element simulation software (ANSYS, ver. 14.5) is used to design and optimize each diaphragm, and the considered sizes of the two beam-supported diaphragms are shown in Table 1.
Table 1. The geometrical parameters of two beam-supported diaphragms.
2.2 Fabrication
The fabrication procedure of the sensing element is illustrated in Fig. 3. To reduce the difficulty in accurately controlling the diaphragm thickness during the fabrication, the beam-supported diaphragm is fabricated on a SOI wafer (device layer 5 µm, box layer 1 µm and handling layer 500 µm) via the MEMS processing technology, which consists of five main steps. (1) Etching the beam structure of the diaphragm in the device layer by using the deep reactive ion etching (DRIE) process [Fig. 3(a)]. (2) Etching a small hole (diameter: 860 µm for 30 kHz diaphragm, 640 µm for 60 kHz diaphragm) in the handling layer also by the DRIE process, and the box layer is used as the self-stopping layer [Fig. 3(b)]. (3) Etching a fiber optic guiding hole (diameter: 950 µm, depth: 150 µm) in the handling layer again by the DRIE process, and moreover, keeping coaxially with the small hole. (4) Completely etching the residual oxide layer beneath the small hole by using the buffered oxide etching (BOE) process to release the beam-supported diaphragm [Fig. 3(d)]. (5) Sputtering a gold film with the thickness of 100 nm onto the beam-supported diaphragm as a reflective surface of Fabry-Perot (FP) cavity [Fig. 3(e)]. After the diaphragm is finished, two beam-supported diaphragms with a natural frequency of 30 kHz and 60 kHz, respectively, are jointly diced from the processed SOI wafer to form a square sensing element of 3.5 mm × 3.5 mm. The microscopic photo of the final sensing element is shown in Fig. 4.
Fig. 3. The schematic of fabrication process for the beam-supported diaphragm: (a) etching the beam structure of the diaphragm in the device layer of a SOI wafer, (b) etching a small hole in the handling layer, (c) etching a fiber optic guiding hole also in the handling layer, (d) removing the oxide layer beneath the small hole, (e) sputtering the gold film on the beam-supported diaphragm, (f) three-dimensional diagram of the beam-supported diaphragm.
Finally, two single-mode fibers (SMF-28e) wrapped by metal sleeve with the diameter of 950 µm are inserted into two fiber optic guiding holes on the sensing element by using an alignment apparatus, which can freely adjust the fiber position in three directions (a 25 nm step in vertical direction and 10 µm in horizontal two directions) to confirm the accuracy of the constructed FP cavity. Once the fiber is adjusted to an optimized working position, the metal sleeve is fixed on the sensing element with the epoxy (353ND), providing a long-term stability. When the fiber is gradually inserted into the fiber optic guiding hole, the reflection spectrums of the FP cavity with different cavity length L is measured, and the results are shown in Fig. 5. The cavity length of the developed sensor is finally set to 150.5 µm since it shows the best contrast approaching one. The final cavity length is larger than the designed etching depth of the fiber guiding hole, the reason of which is that the fabrication accuracy of DRIE process is difficult to be one hundred percent assured. It further illustrates the necessity of using the device layer of SOI wafer to reduce the difficulty in accurately controlling the diaphragm thickness during the fabrication. The assembled sensing element is shown in Fig. 6(a), and it is eventually protected in a printed shell as shown in Fig. 6(b).
Fig. 5. The measured interference spectrum of Fabry-Perot (FP) cavity with different cavity length L.
3. Experimental validation
The schematic diagram of the experimental setup for the whole testing system are presented in Fig. 7, which is mainly composed of 1550 nm narrow linewidth laser [stability: <1% (1 h)], 1 × 2 optic fiber splitter, optic fiber circulator (OFC), photodetector array (COSC, TPIN-LW-M), oscilloscope (Tektronix, MDO3024), ultrasonic loudspeaker (Core morrow), computer and reference microphone (BSWA, MK401-MV401). The original input light emitted by the laser is divided into two equal parts, and each part reaches SMFs through the ports 1 and 2 of two OFCs. Two SMFs are assembled with the two different beam-supported diaphragms on the sensing element, thus forming two FP cavities. Afterwards, the optical signals modulated by both cavities are reflected back into a photodetector array (including two photodetectors) through the ports 2 and 3 of OFCs. Finally, the periodically modulated signal is collected by the oscilloscope, and is processed in a computer. To apply the same ultrasonic pressure to the reference sensor and EFPI sensor, they are placed in a symmetrical position relative to the ultrasonic speaker.
For the beam-supported diaphragm with the designed frequency of 30 kHz, when the output frequency of ultrasonic speaker is 31 kHz, its responsive sensitivity and signal-noise ratio (SNR) is highest, and the values are −15 dB re. 1 V/Pa and 52 dB, respectively, as illustrated in Fig. 8(a). The deviation between the measured and designed resonant frequencies might be caused by the slight fabrication differences in geometrical dimensions. The applied ultrasonic prssure is 108 dB, and thus the noise-limited MDUP of 251 µPa/Hz1/2 can be acquired. When the ultrasonic wave with a frequency of 62 kHz is applied, the time domain waveform and its Fourier transform of the sensor are shown in Fig. 8(b), and its responsive sensitivity and SNR are −38 dB re. 1 V/Pa and 38 dB, respectively. As a result, a 23 dB decrease is observed.
The sensitivity of EFPI sensing system ST can be given by
where R is the responsive sensitivity for the photodetector, and S0 is the FP interference sensitivity. In this research, all the components used for developing the sensing system are selected to operate in the linear region, and S0 and R are constant. Therefore, ST is proportional to Sd. For the sensing diaphragm with the designed frequency of 30 kHz, when the frequency of ultrasonic wave to excite the diaphragm is twice the natural frequency of diaphragm, the responsive sensitivity decreases about 23 dB compared with the resonant sensitivity as mentioned above. So the damping ratio about 0.1 can be acquired according to Eq. (1). Similarly, for the beam-supported diaphragm with the designed frequency of 60 kHz, the resonant frequency of 63 kHz can be obtained, and its time domain waveform and its Fourier transform are shown in Fig. 9(a). The resonant sensitivity and SNR are −22 dB re. 1 V/Pa and 50 dB, respectively. The applied ultrasonic pressure is also 108 dB, and thus the noise-limited MDUP is 316 µPa/Hz1/2. To estimate the damping ratio, the ultrasonic wave with the frequency of 31 kHz is also applied to the diaphragm, and the results are shown in Fig. 9(b). The responsive sensitivity and SNR are −38 dB re. 1 V/Pa and 37 dB, respectively. As a result, a 16 dB decrease is existing. On the basis of Eq. (1), the damping ratio of diaphragm with the designed frequency of 60 kHz comes about 0.07. Figure 10 shows the fitting curve for the sensitivity of the developed sensor versus ultrasonic frequency. When the ultrasonic frequency is lower than 46 kHz, the sensitiviy of 30 kHz diaphragm is higher, or else that of 60 kHz diaphragm is higher. Therefore, the sensor with two resonant frequencies can give a higher sensitivity at different ultrasonic frequencies compared with the approach with only one resonant frequency.Fig. 8. The experimental results of the beam-supported diaphragm with the designed frequency of 30 kHz under: (a) 31 kHz ultrasonic wave, and (b) 62 kHz ultrasonic wave.
Fig. 9. The experimental results of the beam-supported diaphragm with the designed frequency of 60 kHz under: (a) 63-kHz-frequency ultrasonic wave, (b) 31-kHz-frequency ultrasonic wave.
To verify the capability of the developed sensor for picking up the weak PDs, the ultrasonic speaker in Fig. 7 is replaced by a pulsed discharge device to simulate PDs in switchgear. Figure 11(a) shows the time domain waveform picked up by the sensing diaphragm with the designed frequency of 60 kHz. while the time domain waveform of its counterpart shows low SNR as shown in Fig. 11(b). The time domain signals are processed by the fast Fourier transform (FFT). The results show that the diaphragm with the designed frequency of 60 kHz picked up the ultrasonic wave with the frequencies of around 58 kHz and 82 kHz, and the diaphragm with the designed frequency of 30 kHz picked up the ultrasonic wave with the frequencies of around 33 kHz and 82 kHz, as shown in Fig. 11(c). In summary, the frequencies of ultasonic wave generated by the pulsed discharge device mainly concentrated at amplitudes around 33 kHz, 58 kHz and 82 kHz. The ultrasonic pressure at the frequency of 82 kHz is very high, which can be picked up by both diaphragms. While the ultrasonic pressures at the frequencies of 33 kHz and 58 kHz is very small, which can only be picked up by the diaprhagm with high frequency response at the nearby frequency. In general, the wider frequency domain information provides a more accurate detection of the actual cause of insulation damage. As a result, the proposed method owning multiple diaphragms with different natural frequencies on the same sensing element is effective to enhance the capability for detecting the weak PDs.
Fig. 11. (a) The time domain waveform of ultrasonic wave picked up by the diaphragm with the designed frequency of 60 kHz, (b) The time domain waveform of ultrasonic wave picked up by the diaphragm with the designed frequency of 30 kHz, (c) the spectrum of ultrasonic signals picked up by two diaphragms.
4. Conclusions
In summary, a fiber optic EFPI ultrasonic sensor with two resonant frequencies for PDs detection in switchgear is demonstrated. The EFPI sensing element consists of two vibrating diaphragms with different natural frequencies. The noise-limited MDUPs reach 251 µPa/Hz1/2 and 316 µPa/Hz1/2 at the resonant frequencies of 30 kHz and 60 kHz, and the sensitivities of two diaphragms at the resonant frequencies increases 23 dB and 16 dB compared with that at twice or half their resonant frequencies. The presented results are mainly a result of relatively small diaphragm thickness and the application of four-beam supported structure in the sensing diaphragm. The manufacturing method of beam-supported diaphragm on a SOI wafer is proven to be effective to reduce the difficulty in controlling the diaphragm thickness, and enhancing the yield rate for the sensing diaphragms with performance as required. In addition, taking advantage of the method demonstrated in this paper, the sensitive element can also integrate more beam-supported diaphragms to cover more resonant frequencies, leading to a wider frequency response and stronger ability for PDs detection, which is of great potential for practical applications.
Funding
China Scholarship Council.
Acknowledgments
We acknowledge Wenli Li and Ali Jammal for polishing the paper.
Disclosures
The authors declare no conflicts of interest.
References
1. W. Wang, N. Wu, Y. Tian, X. Wang, C. Niezrecki, and J. Chen, “Optical pressure/acoustic sensor with precise Fabry-Perot cavity length control using angle polished fiber,” Opt. Express 17(19), 16613–16618 (2009). [CrossRef]
2. J. Posada-Roman, J. A. Garcia-Souto, and J. Rubio-Serrano, “Fiber optic sensor for acoustic detection of partial discharges in oil-paper insulated electrical systems,” Sensors 12(4), 4793–4802 (2012). [CrossRef]
3. L. Song, Z. Wang, A. Wang, Y. Liu, and K. L. Cooper, “Angular dependence of the frequency response of an extrinsic Fabry-Perot interferometric (EFPI) fiber acoustic sensor for partial discharge detection,” J. Lightwave Technol. 24(9), 3433–3438 (2006). [CrossRef]
4. X. Wang, B. Li, Z. Xiao, S. H. Lee, H. Roman, O. L. Russo, and K. R. Farmer, “An ultra-sensitive optical MEMS sensor for partial discharge detection,” J. Micromech. Microeng. 15(3), 521–527 (2005). [CrossRef]
5. B. Yu, D. W. Kim, J. Deng, H. Xiao, and A. Wang, “Fiber Fabry-Perot sensors for detection of partial discharges in power transformers,” Appl. Opt. 42(16), 3241–3250 (2003). [CrossRef]
6. Y. Huang, T. Jin, and H. Xu, “Research Progress on F-P Interference—Based Fiber-Optic Sensors,” Sensors 16(9), 1424–1438 (2016). [CrossRef]
7. Y. Li, D. Chen, J. Zhang, N. Chen, and X. Gong, “Study of new method combined ultra-high frequency (UHF) method and ultrasonic method on PD detection for GIS,” IOP Conf. Series: Earth and Environmental Science, 94 (012086), 2017.
8. M. MacAlpine, Z. Zhiqiang, and M. S. Demokan, “Development of a fibre-optic sensor for partial discharges in oil-filled power transformers,” Electr. Power Syst. Res. 63(1), 27–36 (2002). [CrossRef]
9. W. Si, C. Fu, D. Li, H. Li, and Y. Yu, “Directional Sensitivity of a MEMS-Based Fiber-Optic Extrinsic Fabry-Perot Ultrasonic Sensor for Partial Discharge Detection,” Sensors 18(6), C1 (2018). [CrossRef]
10. Z. Gong, K. Chen, X. Zhou, Y. Yang, Z. Zhao, H. Zou, and Q. Yu, “High-sensitivity Fabry-Perot interferometric acoustic sensor for low-frequency acoustic pressure detections,” J. Lightwave Technol. 35(24), 5276–5279 (2017). [CrossRef]
11. J. Xia, S. Xiong, F. Wang, and H. Luo, “Wavelength-switched phase interrogator for extrinsic Fabry–Perot interferometric sensors,” Opt. Lett. 41(13), 3082–3085 (2016). [CrossRef]
12. C. Pang, H. Bae, A. Gupta, K. Bryden, and M. Yu, “MEMS Fabry-Perot sensor interrogated by optical system-on-a-chip for simultaneous pressure and temperature sensing,” Opt. Express 21(19), 21829–21839 (2013). [CrossRef]
13. S. Pevec and D. Đonlagić, “Miniature all-fiber Fabry-Perot sensor for simultaneous measurement of pressure and temperature,” Appl. Opt. 51(19), 4536–4541 (2012). [CrossRef]
14. O. C. Akkaya, O. Kilic, M. J. F. Digonnet, G. S. Kino, and O. Solgaard, “Modeling and Demonstration of Thermally Stable High-Sensitivity Reproducible Acoustic Sensors,” J. Microelectromech. Syst. 21(6), 1347–1356 (2012). [CrossRef]
15. J. Ma, H. Xuan, H. Lut Ho, W. Jin, Y. Yang, and S. Fan, “Fiber-Optic Fabry–Pérot Acoustic Sensor With Multilayer Graphene Diaphragm,” IEEE Photonics Technol. Lett. 25(10), 932–935 (2013). [CrossRef]
16. B. Liu, J. Lin, H. Liu, A. Jin, and P. Jin, “Extrinsic Fabry-Perot fiber acoustic pressure sensor based on large-area silver diaphragm,” Microelectron. Eng. 166(6), 50–54 (2016). [CrossRef]
17. W. Jo, O. C. Akkaya, O. Solgaard, and M. J. F. Digonnet, “Miniature fiber acoustic sensors using a photonic-crystal diaphragm,” Opt. Fiber Technol. 19(6), 785–792 (2013). [CrossRef]
18. Q. Chen, W. Zhang, and H. Zhao, “Response Bandwidth Design of Fabry-Perot Sensors for Partial Discharge Detection Based on Frequency Analysis,” J. Sens. 2019(6), 1–11 (2019). [CrossRef]
19. C. Fu, W. Si, H. Li, D. Li, P. Yuan, and Y. Yu, “A novel high-performance beam-supported membrane structure with enhanced design flexibility for partial discharge detection,” Sensors 17(3), 593–604 (2017). [CrossRef]
