Laser & Optoelectronics Progress, Volume. 61, Issue 17, 1700005(2024)
Research Progress on Improving the Temperature Measurement Accuracy of Raman Distributed Fiber Sensing
Figures & Tables(18)
Fig. 1. Experimental setup based on dual-channel temperature demodulation technology
Fig. 3. Experimental setup based on single-channel temperature demodulation technology
Fig. 5. Experimental set-up of scheme for compensating light attenuation with dual laser pumping and Raman OTDR traces correspondingly. (a) Experimental set-up; (b) Raman OTDR traces
Fig. 6. Experimental set-up of optical attenuation calibration scheme of entire distributed fiber and Raman OTDR traces correspondingly. (a) Experimental set-up; (b) Raman OTDR traces
Fig. 7. Experimental set-up of self-calibration scheme of end-reflection of fiber and Raman OTDR traces correspondingly. (a) Experimental set-up; (b) Raman OTDR traces
Fig. 8. Experimental set-up of double-end loop demodulation scheme and Raman OTDR traces correspondingly. (a) Experimental set-up; (b) Raman OTDR traces
Fig. 10. Demodulation scheme diagram. (a) Traditional demodulation scheme; (b) dynamic difference attenuation recognition scheme
Fig. 11. Temperature measurement accuracy at different sensing distances[101]. (a1) Experimental result of traditional demodulation scheme at 1.38 km; (b1) experimental result of traditional demodulation scheme at 10.22 km; (c1) experimental result of traditional demodulation scheme at 18.80 km; (d1) experimental result of traditional demodulation scheme at 28.90 km; (a2) experimental result of dynamic gain calibration scheme at 1.38 km; (b2) experimental result of dynamic gain calibration scheme at 10.22 km; (c2) experimental result of dynamic gain calibration scheme at 18.80 km; (d2) experimental result of dynamic gain calibration scheme at 28.90 km
Fig. 12. Temperature measurement results at a centimeter-level spatial scale detected using a conventional demodulation method and slope-assisted demodulation method[53]. (a) Temperature measurement results measured using traditional demodulation method; (b) distribution of superimposed Raman OTDR trace after attenuation compensation; (c) temperature measurement results measured using slope-assisted coefficient demodulation method
Table 1. Comparison and analysis of different Raman light attenuation compensation schemes
View tableTable 1. Comparison and analysis of different Raman light attenuation compensation schemes
Scheme Type Function Advantage Characteristic Representative Result Dual laser pumping Compensation for difference in optical attenuation between Raman Stokes and anti-Stokes signals Omits pre-calibration process and eliminates effects of fiber dispersion An additional cost and complex structure for practical applications［ 67 ］‒ Optical attenuation calibration of the sensing fiber Compensation for attenuation varying with sensing distance Operation is simple without additional devices Re-calibration is required after replacing any equipment or sensing fiber［
59 ］Sudden fiber loss cannot be erased
0.36 ℃ at 10.00 km［
59 ］Figure of merit：27.77
End reflection self-calibration Compensation for attenuation varying with time Eliminates additional fiber loss caused by fiber bending SNR is lower than single-ended and double-ended demodulation［
64 -65 ］Data processing is more difficult［
67 ］2.95 ℃ at 4.20 km［
61 ］Figure of merit：1.42
Double-end loop demodulation Avoid effects of fiber bending on the measurement results
More appropriate to engineering applications
Requires twice length of sensing fiber and longer measurement time［
67 ］The SNR is the best at the middle of the sensing fiber
1.10 ℃ at 19.00 km［
64 ］Figure of merit：17.27
0.43 ℃ at 10.00 km［
65 ］Figure of merit：23.25
Attenuation difference fitting Eliminates additional fiber loss caused by fiber bending Sudden fiber loss cannot be erased 1.56 ℃ at 10.00 km［
67 ］Figure of merit：6.41
Table 2. Typical pulse coding schemes and experimental results
View tableTable 2. Typical pulse coding schemes and experimental results
Scheme Sensing distance /km Temperature accuracy /℃ Figure of merit Low-Repetition-Rate cyclic pulse coding［ 71 ］26.00 3.00 8.66 255-bit simplex coded［ 72 ］19.50 3.00 6.50 Cyclic pseudo-random sequence coding［ 73 ］1.00 1.50 0.66 Pre-Shaped simplex code［ 74 ］50.00 9.00 5.55 Improved simplex coding［ 75 ］1.00 0.10 10.00 Genetic-Optimized aperiodic code［ 76 ］39.00 3.90 10.00 Golay correlation sequence coding［ 77 ］45.00 2.50 18.00 3-bit simplex coded［ 78 ］9.01 0.50 18.02
Table 3. Typical algorithm denoising schemes and experimental results
View tableTable 3. Typical algorithm denoising schemes and experimental results
Scheme Sensing distance /km Temperature accuracy /℃ Figure of merit Short term Fourier transform［ 80 ］0.23 0.50 0.46 Dynamic sampling-correction scheme［ 81 ］20.00 1.00 20.00 Deep one-dimensional denoising Convolutional neural network［ 86 ］10.00 0.70 14.28 Deconvolution algorithms［ 88 ］30.00 2.70 11.11 Total variation deconvolution［ 90 ］1.30 0.10 13.00 D-SVD algorithms［ 91 ］0.60 0.87 0.68 Principal component analysis［ 92 ］0.65 0.30 2.16
Table 4. Typical Rayleigh noise filtering schemes and experimental results
View tableTable 4. Typical Rayleigh noise filtering schemes and experimental results
Scheme Sensing distance /km Temperature accuracy /℃ Figure of merit Data fitting scheme［ 93 ］1.60 0.50 3.20 Adaptive temperature demodulation scheme［ 94 ］7.00 1.00 7.00 Dual fiber calibration scheme［ 95 ］9.10 1.70 5.35
Table 5. Comparison and analysis of different SNR improvement schemes
View tableTable 5. Comparison and analysis of different SNR improvement schemes
Scheme Type Advantage Characteristic Representative Result Pulse Coding It can effectively improve SNR and sensing distance simultaneously
Denoising result is better than other denoising methods
Additional costs and complexity in practical application 3.90 ℃ at 39.00 km［
76 ］Figure of merit：10.00
6.20 ℃ at 1.00 km［
75 ］Figure of merit：10.00
Algorithm Denoising Operation process is simple and reduce average number of times and optimize measurement time of system Some useful signals may be filtered out 2.70 ℃ at 30.00 km［
85 ］Figure of merit：11.11
Rayleigh Optical Noise Suppression Operation process is simple It cannot eliminate electrical noise，and needs to be used in conjunction with other denoising methods 1.70 ℃ at 9.10 km［
95 ］Figure of merit：5.35
Noise Suppression in the Demodulation Stage Calibration process is eliminated，and operation is convenient It needs to be used in conjunction with other denoising methods 0.18 ℃ at 17.00 km［
97 ］Figure of merit：94.40
Table 6. Comparison and analysis of different improved Raman transfer equation schemes
View tableTable 6. Comparison and analysis of different improved Raman transfer equation schemes
Scheme Type Advantage Characteristic Representative Result Compensation for fiber dispersion No additional equipment is required for operation Use multimode sensing fiber to compensate for fiber dispersion 1.20 ℃ at 13.00 km［
59 ］Figure of merit：10.83
Compensation for temperature sensitivity of APD Realize Raman distributed temperature sensing over long sensing distance Additional costs in practical application 7.20 °C at 30.00 km［
98 ］Figure of merit：4.16
2.10 °C at 18.80 km［
98 ］Figure of merit：8.95
Compensation for temperature Sensitivity of sensing fiber It can improve temperature sensitivity of sensing fiber at different distances Additional complexity in practical application 0.36 °C at 10.00 km［
100 ］Figure of merit：27.78
Tools
Get Citation
Copy Citation Text
Kangyi Cao, Jian Li, Mingjiang Zhang. Research Progress on Improving the Temperature Measurement Accuracy of Raman Distributed Fiber Sensing[J]. Laser & Optoelectronics Progress, 2024, 61(17): 1700005
Paper Information
Category: Reviews
Received: Oct. 24, 2023
Accepted: Dec. 18, 2023
Published Online: Sep. 14, 2024
The Author Email: Jian Li (lijian02@tyut.edu.cn), Mingjiang Zhang (zhangmingjiang@tyut.edu.cn)
CSTR:32186.14.LOP232358
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20