CN111121836A - Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting - Google Patents
Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting Download PDFInfo
- Publication number
- CN111121836A CN111121836A CN201911309627.4A CN201911309627A CN111121836A CN 111121836 A CN111121836 A CN 111121836A CN 201911309627 A CN201911309627 A CN 201911309627A CN 111121836 A CN111121836 A CN 111121836A
- Authority
- CN
- China
- Prior art keywords
- brillouin
- frequency shift
- spectrum
- processing method
- brillouin frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000605 extraction Methods 0.000 title claims abstract description 31
- 238000003672 processing method Methods 0.000 claims abstract description 109
- 238000001228 spectrum Methods 0.000 claims abstract description 86
- 238000005259 measurement Methods 0.000 claims abstract description 28
- 238000000034 method Methods 0.000 claims abstract description 20
- 239000013307 optical fiber Substances 0.000 claims abstract description 17
- 230000003595 spectral effect Effects 0.000 claims description 14
- 230000008859 change Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 abstract description 42
- 238000013459 approach Methods 0.000 abstract description 3
- 238000004458 analytical method Methods 0.000 description 8
- 239000000835 fiber Substances 0.000 description 5
- 238000010408 sweeping Methods 0.000 description 5
- 238000011160 research Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N Silicium dioxide Chemical compound O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000010183 spectrum analysis Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 241000544061 Cuculus canorus Species 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010292 electrical insulation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000002657 fibrous material Substances 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000000377 silicon dioxide Substances 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D5/00—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable
- G01D5/26—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light
- G01D5/32—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light
- G01D5/34—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells
- G01D5/353—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells influencing the transmission properties of an optical fibre
- G01D5/35338—Mechanical means for transferring the output of a sensing member; Means for converting the output of a sensing member to another variable where the form or nature of the sensing member does not constrain the means for converting; Transducers not specially adapted for a specific variable characterised by optical transfer means, i.e. using infrared, visible, or ultraviolet light with attenuation or whole or partial obturation of beams of light the beams of light being detected by photocells influencing the transmission properties of an optical fibre using other arrangements than interferometer arrangements
- G01D5/35354—Sensor working in reflection
- G01D5/35358—Sensor working in reflection using backscattering to detect the measured quantity
- G01D5/35364—Sensor working in reflection using backscattering to detect the measured quantity using inelastic backscattering to detect the measured quantity, e.g. using Brillouin or Raman backscattering
Abstract
A Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting is characterized in that in a measurement signal output by a distributed optical fiber sensor, the Brillouin frequency shift is used as the center to intercept the line width delta v of 1 Brillouin gain spectrumBFitting the signal by a quadratic polynomial processing method to obtain a Brillouin spectrum: gB(v)＝av2+ bv + c, the Brillouin frequency shift is
Description
Technical Field
The invention relates to an improved Brillouin frequency shift rapid extraction method, and belongs to the technical field of measurement.
Background
The distributed optical fiber sensor has the advantages of electromagnetic interference resistance, corrosion resistance, good electrical insulation property and the like of a common optical fiber sensor, and also has the unique advantages of acquiring the distribution information of the measured field along the whole optical fiber by one-time measurement, and the like. Therefore, the application field is very wide. Among them, the distributed sensing technology based on fiber brillouin scattering has higher measurement accuracy, measurement range and spatial resolution than other types of distributed fiber sensing technologies in temperature and strain measurement, and thus has attracted extensive attention and research at home and abroad. The technology is used for carrying out online monitoring on the temperature and the strain of an oil-gas pipeline, a large-scale water conservancy and hydropower engineering structure, an electric power cable and the like, and the quick and accurate positioning of the hidden trouble and the fault point can be realized.
The Brillouin frequency shift is in a linear relation with temperature and strain and is the most stable among spectral characteristic quantities related to the temperature and the strain, and at present, Brillouin frequency shift information needs to be extracted in most optical fiber distributed sensing measurement based on Brillouin scattering. The accurate and rapid measurement of Brillouin frequency shift is very critical to the optical fiber distributed sensing system based on Brillouin scattering. In general, the brillouin frequency shift extraction method is mainly based on a fitting-based method, and can be divided into a model-based processing method and a non-model-based processing method. Because the Brillouin frequency shift can be extracted by fully utilizing waveform data of the Brillouin spectrum, the overall processing method based on the model has higher accuracy. When the width of an incident rectangular pulse is obviously larger than 10ns (larger than 50ns), the measured Brillouin spectrum in the optical fiber approximately meets the Lorentz function, and the fitting processing method based on the Lorentz model is most widely applied at present. Of course, as the spatial resolution is reduced with the increase of the pulse width, a smaller pulse width is often used to improve the spatial resolution, and when the value of the pulse width is less than 10ns, the brillouin spectrum approximately satisfies the gaussian distribution, and in this case, the processing method based on the gaussian model should be adopted. Obviously, the mode of selecting the processing method according to the pulse width adds an interference factor to the extraction of the brillouin frequency shift, and a processing method based on a general model is urgently needed. The Brillouin spectrum can be considered to better satisfy the Voigt function in the whole pulse width range of the rectangular pulse, so that the processing method based on the Voigt model has higher accuracy. However, the Voigt function is not an algebraic expression, only has a numerical solution, and although the processing method based on the Voigt function has the highest accuracy, the calculation speed is too slow, so that the Voigt function is not particularly suitable for occasions with a long sensing range and high spatial resolution. In order to improve the calculation speed, a pseudo Voigt model linearly combined by Lorentz and Gaussian models is adopted to approximate the Brillouin spectrum, and the method has a good characterization effect. To further improve the calculation speed and convergence ability, a lot of work has been done. In order to improve the accuracy of Brillouin frequency shift extraction, some scholars adopt finite element analysis to improve the Newton processing method and the Levenberg-Marquardt processing method, and then adopt the calculation results of Brillouin spectra with different signal-to-noise ratios and line widths to verify the effectiveness of the improved processing method in the aspect of precision improvement. Aiming at the problem that the divergence of the processing method is possibly caused by the difficulty in the selection of the initial variable value, some scholars respectively adopt cuckoo Newton search and particle swarm optimization to obtain the initial value closer to the optimal solution, and then adopt a Levenberg-Marquardt processing method to optimize the variable, so that the convergence can be ensured. Although the accuracy is improved and the divergence probability of the processing method is reduced by the processing method, the model corresponds to a nonlinear objective function, an iterative optimization processing method is needed to obtain an optimal solution, and the calculation speed is still slow under normal conditions. This is more prominent when the sensing distance is longer and the spatial resolution is higher. This problem remains to be investigated further.
Disclosure of Invention
The invention aims to provide a Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting, aiming at overcoming the defects in the prior art, and improving the processing speed while ensuring the extraction precision of the Brillouin frequency shift.
The problems of the invention are solved by the following technical scheme:
a Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting is characterized in that aiming at a measurement signal output by a distributed optical fiber sensor, the Brillouin frequency shift is used as the center to intercept the line width delta v of 1 Brillouin gain spectrumBFitting the signal by a quadratic polynomial processing method to obtain a Brillouin spectrum:
gB(v)＝av2+bv+c
wherein a, b and c are second-order polynomial coefficients, the numerical values of which are different along with the change of the Brillouin spectrum, and the Brillouin frequency shift is
The Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting intercepts the line width delta v of 1 Brillouin gain spectrum by taking the Brillouin frequency shift as the centerBThe brillouin frequency shift is replaced by the frequency corresponding to the brillouin spectral peak.
The method selects a Brillouin spectrum with a symmetrical line width near a peak value, then adopts a quadratic polynomial to approach the Brillouin spectrum so as to extract Brillouin frequency shift, and not only can greatly improve the calculation speed, but also the calculation time is only 1.15%, 1.80%, 1.51% and 0.51% of the existing typical method (a fitting method based on Lorentz, Gaussian, pseudo Voigt and Voigt models). And the calculation accuracy is equivalent to the typical processing method based on the Lorentz, Gaussian, pseudo Voigt and Voigt model based fitting method, and the difference of the temperature is converted into the temperature errors of only 0.35 ℃, 0.21 ℃ and 0.23 ℃. The effectiveness of the processing method is verified by adopting numerical simulation and actual measurement of the Brillouin spectrum. The invention provides good support for realizing the rapid measurement of the optical fiber distributed sensing.
Drawings
The present invention will be described in further detail with reference to the accompanying drawings.
FIG. 1 is a fitting result based on pseudo Voigt and quadratic polynomial processing methods;
FIG. 2 shows the result of extraction of Brillouin frequency shift and the calculation time of the Brillouin frequency shift, actually measured spectrum, in different processing methods, wherein FIG. 2(a) shows the Brillouin frequency shift, and FIG. 2(b) shows the calculation time;
FIG. 3 is a fitting result of Brillouin spectra of different processing methods, actually measured spectra;
FIG. 4 is a simulation spectrum of the results of Brillouin frequency shift extraction for different processing methods;
FIG. 5 is a fitting result, simulated spectrum, of Brillouin spectra for different processing methods;
FIG. 6 is a relationship between Brillouin frequency shift error and sweep range;
FIG. 7 is a relationship between Brillouin frequency shift error and frequency sweep point number;
FIG. 8 is a Brillouin frequency shift error versus signal-to-noise ratio;
FIG. 9 is a Brillouin frequency shift error versus linewidth;
FIG. 10 is a relationship between Brillouin frequency shift error and deviation of frequency sweep range;
FIG. 11 shows the result of Brillouin frequency shift extraction and the calculation time of the conventional processing method, and the actually measured spectrum, wherein FIG. 11(a) shows the Brillouin frequency shift, and FIG. 11(b) shows the calculation time;
fig. 12 shows a simulated spectrum of the brillouin shift extraction result of the present invention and the classical processing method.
The symbols used herein are: v is the frequency value; v. ofBIs a brillouin frequency shift; Δ vBLAnd Δ vBGLine widths of a Lorentzian type Brillouin spectrum and a Gaussian type Brillouin spectrum are respectively set; Δ vBSetting the line width of the Brillouin gain spectrum; g01And g02Respectively Lorentzian and Gaussian peak gains of the Brillouin scattering spectrum; evBThe average value of the Brillouin frequency shift error amplitude is obtained; gB(v) Is the gain of the brillouin scattering spectrum; v is frequency; a, b and c are coefficients of a quadratic polynomial.
Detailed Description
Aiming at the calculation of actual measurement and simulation Brillouin spectrum, the calculation speed of the quadratic polynomial fitting processing method is obviously improved compared with the existing classical model based on the processing method, but the error is larger. The system researches the influence of the frequency sweep range, the number of frequency sweep points, the signal-to-noise ratio, the line width and the deviation of the frequency sweep range on the extraction accuracy of the Brillouin frequency shift based on the quadratic polynomial. According to research results, the invention provides an improved quadratic polynomial fitting processing method, and the improved processing method not only can greatly improve the calculation speed, but also has the calculation accuracy equivalent to that of a classical processing method. The effectiveness of the processing method is verified by adopting numerical value generation and actual measurement of the Brillouin spectrum.
1 principle of treatment method
Brillouin scattering occurs when the incident pulsed light is rectangular wave, and the Brillouin spectrum approximately meets the following Voigt model which is a convolution form of Lorentz function and Gaussian function
Wherein v is a frequency value in GHz; v. ofBIs Brillouin frequency shift, with unit being GHz; Δ vBLAnd Δ vBGThe linewidths of the Lorentzian type and Gaussian type brillouin spectra are respectively, and the unit is GHz. The Brillouin spectrum can be represented by adopting a pseudo Voigt model with faster calculation, and the line width of the Brillouin gain spectrum is set to be delta vB；g01And g02Lorentzian and Gaussian peak gains, respectively, of the Brillouin scattering spectrum, which is expressed as follows
Is provided with
v＝x+vB(3)
When formula (3) is substituted for formula (2) and expanded according to Taylor series
Where O (x) is the higher order infinitesimal of x.
If x is small, e.g., x ≦ Δ vBNeglecting the high order infinitesimal magnitude, there is a
According to formulae (3) and (5) there are
Thus, the brillouin gain spectrum approximately satisfies a quadratic polynomial relationship over a range of line widths. The Brillouin spectrum can be approximated using a quadratic polynomial, i.e.
gB(v)＝av2+bv+c (7)
The Brillouin frequency shift calculation formula after fitting by using quadratic polynomial is as follows
Note that if the brillouin frequency shift is determined to be outside the sweep range of the spectrum, the brillouin frequency shift error may be larger than normal. And selecting random values which meet uniform distribution in a sweep frequency range as the calculated Brillouin frequency shift when the processing method is implemented, so that the measurement error can be reduced, and the measurement result is more normal.
As can be seen from fig. 1, the pseudo Voigt function can well approximate the Voigt function, and the quadratic polynomial is expected to approximate the corresponding spectral line although it is not slightly different from the real Voigt function in terms of expression.
2 processing methods comparison
2.1 actual measurement spectra
The invention builds an Optical fiber Brillouin spectrum measuring system based on AV6419 type Optical Time Domain reflectometer (BOTDR) produced by the middle electrical and electronic instruments and meters Limited company, and selects SM 9/125 μm Optical fiber of about 1 km. The sweep frequency range is 10.52-10.92 GHz, the sweep frequency interval is 1MHz, the wavelength of incident pulse light is 1550nm, the pulse width is 10ns, the sampling resolution is 10m, and the superposition average frequency is 218. The experiment was performed at room temperature, but the brillouin shift along the line was not constant due to the strain experienced by the wound fiber. The method is calculated by a model processing method based on Lorentz, Gaussian, pseudo Voigt and a quadratic polynomial fitting processing method, and the extraction result and the calculation time of the Brillouin frequency shift are shown in figure 2. The fitting results of 5 processing methods of a typical spectrum were chosen as shown in fig. 3.
As can be seen from fig. 2(a), the brillouin frequency shift extraction results based on the processing methods of the lorentz, gaussian, pseudo Voigt and Voigt models are almost consistent, which is consistent with the result that the fitting results based on the processing methods of the gaussian, pseudo Voigt and Voigt models in fig. 3 can better approach the actually measured spectrum, and the results also basically verify the reliability of the brillouin frequency shift extraction based on the spectrum models. However, the brillouin extraction result based on the quadratic polynomial processing method is smaller than the calculation results of other processing methods in most regions, and the average values of the differences from the above 4 processing methods are 0.70, 0.74, 0.73 and 0.73MHz, respectively, and the maximum values of the differences are 2.47, 2.60, 2.57 and 2.60MHz, respectively. Since the above analysis considers that the calculation result based on the spectral model processing method is reliable, there is a significant error in the quadratic polynomial fitting processing method, which is consistent with the comparison of the quadratic polynomial fitting result with the large difference of the measured spectrum (fig. 3). This conclusion will of course be further verified in subsequent analyses. As can be seen from fig. 2(b), the calculation speed is slowest based on the Voigt model processing method, then the processing methods based on the lorentz, pseudo-Voigt and gaussian models are sequentially performed, the fastest is the quadratic polynomial fitting processing method, the average calculation times corresponding to the 5 processing methods in fig. 2(b) are 58.97ms, 37.69ms, 44.72ms, 131.48ms and 6.76ms, respectively, that is, the calculation time of the quadratic polynomial fitting processing method is only 1.15%, 1.80%, 1.51% and 0.51% of the previous 4 processing methods. That is to say, the quadratic polynomial fitting processing method has a very fast calculation speed, but the calculation error may be too large, so that the processing method needs to be corrected, which is the core of the technical scheme of the present invention.
Note that the present invention does not analyze the processing method accuracy from fiber brillouin spectral data at temperature and strain. This is because this method is usually not suitable because it is assumed that the processing method is accurate, and the accuracy of the quadratic polynomial fitting processing method itself needs to be checked. The reason for selecting a signal-to-noise ratio signal in the present invention is that excessive noise may mask the law of the brillouin frequency shift with the position of the optical fiber.
2.2 numerical production of spectra
2.1 section calculation of the existing measured spectrum, the reason for further calculation of the spectrum generated by adopting numerical values in this section is as follows: 1) although the measured spectrum is reliable, Brillouin frequency shift cannot be obtained accurately enough; 2) subsequent analysis involving large numbers of numerically generated spectra requiresThe reliability of the spectral analysis results generated by the verification values. The Brillouin spectrum is generated according to equation (1), A, v thereinB、ΔvBLAnd Δ vBGObtained by calculation of actual measurement spectrum of section 2.1 by a Voigt model processing method. The signal-to-noise ratio is consistent with the measured spectrum, about 33 dB. The extraction result of the brillouin frequency shift is shown in fig. 4. The results of the 5 processing methods are shown in FIG. 5 for a typical spectrum fit to the measured spectrum of FIG. 3.
As can be seen from fig. 4, the brillouin frequency shift calculated by the 5 processing methods and the change rule with the position of the optical fiber are very close to the actually measured spectrum. Still, the results of brillouin frequency shift extraction based on the lorentz, gaussian, pseudo Voigt and Voigt model processing methods are almost consistent, and the calculation result of the quadratic polynomial fitting processing method is obviously smaller than that of the quadratic polynomial fitting processing method, which is very consistent with fig. 2. Comparing fig. 3 and 5, it can be seen that the fitting results of the measured spectrum and the simulated spectrum are very consistent. This verifies that the numerically generated spectrum signal can better simulate the measured spectrum. Since the brillouin frequency shift of the numerically generated signal is known, the mean values of the error amplitudes for the 5 processing methods are known to be 0.06, 0.05, and 1.35MHz, respectively. This is basically consistent with the results based on the actually measured spectrum analysis, i.e. the accuracy of the processing method based on the lorentz, gaussian, pseudo-Voigt and Voigt models is high enough, but significant errors exist in the quadratic polynomial fitting processing method. The reliability of the numerical generation spectrum is also basically verified.
3. Influencing factors of polynomial fitting method
In order to reduce the error of the quadratic polynomial fitting processing method, the influence rule of various factors on the accuracy of the processing method needs to be researched. According to system research, the influence of spectrum model selection on the Brillouin frequency shift extraction result is small when a signal is generated by a numerical value, and meanwhile, in order to properly accelerate the calculation speed, a Lorentz model is adopted for generating a spectrum in subsequent analysis. Considering that the actual conditions are variable, the value range of the parameters is properly expanded compared with the actual conditions during analysis, and the reliability of the method is not reduced.
3.1 sweep Range
g0、vBAnd Δ vBRespectively taking 0.9 GHz, 10.7GHz and 0.03 GHz; the signal-to-noise ratios were set at 10, 20 and 30dB(ii) a The number of frequency sweeping points is 61; the sweep frequency range is 0.2 delta vBTo 10 Δ vBWithin a range. 10000 spectrum signals are generated according to each parameter combination, and the average value of the Brillouin frequency shift error amplitude is obtained after calculation. The brillouin frequency shift error calculated by the quadratic polynomial fitting processing method is shown in fig. 6.
As can be seen from fig. 6, when the frequency sweep point number is fixed, the calculation error of brillouin frequency shift is large when the frequency sweep range is small. The error is gradually reduced along with the increase of the sweep frequency range, and the error reaches the minimum value when the sweep frequency range is one line width. Then, the error gradually increases again as the sweep range increases. This is because the corresponding signal does not contain enough spectral features when the sweep range is too small, and the number of points in the spectral feature extraction effective range is too small when the sweep range is too large. The error is therefore large in both cases.
3.2 frequency sweep points
Sweep range of Δ vB(ii) a The number of sweep points N varies from 3 to 501. Other parameters are consistent with section 3.1, and the brillouin frequency shift error calculated by the quadratic polynomial fitting processing method is shown in fig. 7.
As can be seen from fig. 7, the brillouin frequency shift error gradually decreases as the number of frequency sweep points increases. Fitting to find the mean value E of the Brillouin frequency shift error amplitudevBSatisfies the following conditions: evB＝aNbIn the above 3 cases, the values of a were 3.2812, 0.8580, and 0.2677, respectively, and the values of corresponding b were-0.5095, -0.4722, and-0.4710, respectively, with the fitting relative errors of 3.11%, 4.13%, and 4.12%, respectively.
3.3 Signal-to-noise ratio
Sweep range of Δ vB(ii) a The signal-to-noise ratio varies in the range of 0 to 40 dB. Other parameters are consistent with section 3.1, and the brillouin frequency shift error calculated by the quadratic polynomial fitting processing method is shown in fig. 8.
As can be seen from fig. 8, the brillouin frequency shift error gradually decreases as the signal-to-noise ratio (SNR) increases. When the signal-to-noise ratio is larger than 10dB, the Brillouin frequency shift error and the signal-to-noise ratio approximately meet the exponential change. When the signal-to-noise ratio is smaller than 10dB, the change rule of the error is different from that when the signal-to-noise ratio is larger than 10dB, because the Brillouin frequency shift obtained by directly adopting a quadratic polynomial fitting processing method exceeds the frequency sweeping range when the signal-to-noise ratio is low, and the random value in the frequency sweeping range is adopted as the calculated Brillouin frequency shift when the processing method is realized, so that the measurement error can be reduced, and the measurement result is more normal.
3.4 line width
Sweep range of Δ vB；ΔvBVarying in the range of 0.03 to 1 GHz. Other parameters are consistent with section 3.1, and the brillouin frequency shift error calculated by the quadratic polynomial fitting processing method is shown in fig. 9.
As can be seen from fig. 9, the brillouin frequency shift error increases linearly with an increase in line width. Therefore, spatial resolution cannot be improved by selecting an excessively narrow incident pulsed light. To improve the accuracy of the brillouin frequency shift, an appropriate configuration should be selected such that the brillouin spectral linewidth is reduced.
3.5 sweep Range bias
Because the Brillouin frequency shift of the optical fiber to be measured cannot be completely and accurately obtained, the midpoint of the frequency sweep range and the Brillouin frequency shift are not necessarily coincident, and the difference between the midpoint of the frequency sweep range and the Brillouin frequency shift is called as frequency sweep range deviation in the method. Deviation of sweep frequency range from 0 to 0.3 delta vB(ii) a variation within a range; sweep range of Δ vB. Other parameters are consistent with section 3.1, and the brillouin frequency shift error calculated by the quadratic polynomial fitting processing method is shown in fig. 10.
As can be seen from fig. 10, the brillouin frequency shift error tends to increase as the midpoint of the sweep range gradually deviates from the brillouin frequency shift. Therefore, the spectral signal should be selected for brillouin frequency shift extraction centered around brillouin frequency shift.
4 improved processing method and verification
4.1 improved treatment
From the above analysis, the factors having a large influence on the accuracy of the quadratic polynomial fitting processing method are the frequency sweep range, the number of frequency sweep points, the signal-to-noise ratio, the line width and the deviation of the frequency sweep range. Although the frequency sweeping point number, the signal-to-noise ratio and the line width have great influence on the accuracy of the quadratic polynomial fitting processing method, the 3 factors are mainly determined by the actual measurement condition. And sweep range bias for fitting spectraThe near optimal value can be adjusted to reduce the error of the processing method. In particular, 1 av is truncated, centered on the brillouin shift, instead of all measurement signalsBThe spectral signal of (a) is used for fitting of a quadratic polynomial processing method. The improved quadratic polynomial fitting processing method has improved accuracy and performance, and the calculation speed of the quadratic polynomial fitting processing method is further accelerated due to the reduction of the number of points to be fitted.
Because the Brillouin frequency shift is to be measured and is unknown before the processing method is executed, 1 delta v is intercepted by taking the corresponding frequency of the Brillouin spectrum peak value as the center during the actual processing methodBThe spectral signal of (a) is used for fitting of a quadratic polynomial processing method. The line width is related to a plurality of factors including the fiber density, the refractive index, the viscosity coefficient of the silica fiber material, the laser output center wavelength, the incident pulse light width, etc., and generally, the experimental arrangement is basically fixed except for the incident pulse light width. Therefore, the estimation of the line width is easy.
4.2 actual measurement spectra
The analysis is carried out by adopting an actually measured spectrum of section 2.1, and the difference is that the section only intercepts 1 delta v by taking a gain peak value as a centerBOf the spectrum of (a). Δ vBAn approximation of the actual measurement is taken, i.e. 0.1 GHz. The calculation results and calculation times of the 5 processing methods are shown in fig. 11.
As can be seen from fig. 11(a), the results of brillouin frequency shift extraction based on the lorentz, gaussian, pseudo Voigt, and Voigt model processing methods almost agree with those of fig. 2 (a). As can be seen from comparing fig. 2(a) and 11(a), the calculation result of the improved quadratic polynomial fitting processing method proposed by the present invention is very close to that of the classical spectral model processing method, and the mean values of the differences from the above 4 processing methods are only 0.39, 0.24, 0.23 and 0.26MHz, respectively. Considering that the temperature coefficient is 1.12 MHz/DEG C, the temperature errors corresponding to the Brillouin frequency shift error above are only 0.35 ℃, 0.21 ℃ and 0.23 ℃ respectively. In addition, differences of Brillouin frequency shift extraction results of the 4 classical processing methods before and after adjustment of a spectrum range to be fitted are small, and the accuracy of the 4 processing methods is further verified. Fig. 11(b) is very similar to fig. 2(b), and the second order polynomial fitting processing method is much less computationally intensive than the other processing methods. The above results illustrate that the improved quadratic polynomial fitting processing method is far less computationally intensive than other classical processing methods, but with accuracy comparable to other processing methods.
4.3 numerical production of spectra
The spectrum parameters and the spectrum generation method are consistent with those in section 2.2, but the section intercepts 0.1GHz spectrum signals by taking a gain peak value as a center and is used for Brillouin frequency shift extraction. The results of the 5 processing methods are shown in fig. 12.
As can be seen from a comparison of fig. 11(a) and 12, the calculation results of the measured spectrum and the simulated spectrum are very similar for the 5 processing methods. Spectra were generated for values whose mean differences from the above 4 treatment methods were only 0.34, 0.21, 0.23 and 0.24MHz, respectively. The mean values of the error amplitudes of the 4 classical treatment methods and the improved treatment method provided by the invention are only 0.13, 0.06 and 0.24MHz respectively, the average values are estimated according to a typical temperature sensitive coefficient of 1.2 MHz/DEG C, and the errors of the 5 treatment methods are 0.11, 0.05 and 0.20 ℃ respectively when the temperature is measured by a single factor. The above results further verify the effectiveness of the improved quadratic polynomial fitting processing method provided by the invention.
5 conclusion
Based on actual measurement and Brillouin spectrum generated by numerical values, the Brillouin frequency shift extraction method based on the quadratic polynomial fitting processing method is systematically processed, so that the calculation accuracy is remarkably improved on the basis of ensuring the real-time property, and particularly, the following conclusions are drawn:
1) when the number of the frequency sweep points is fixed, the Brillouin frequency shift error is reduced to the minimum value and then gradually increased along with the increase of the frequency sweep range, and the optimal frequency sweep range is 1 line width; increasing Brillouin frequency shift errors along with the number of frequency sweeping points and the signal-to-noise ratio to respectively form power and exponential rules to reduce; the error linearly increases with the increase of the line width; the error is gradually increased along with the increase of the deviation of the sweep frequency range, and the spectrum signals for feature extraction are selected to be symmetrical around Brillouin frequency shift as much as possible.
2) The calculation speed of the quadratic polynomial fitting processing method is far faster than that of the processing method based on Lorentz, Gaussian, pseudo Voigt and Voigt models,but the original quadratic polynomial fitting process may have significant errors. Improved quadratic polynomial fitting processing method (1 Deltav is intercepted by taking corresponding frequency of Brillouin spectrum peak as center) through selecting proper sweep frequency rangeBThe spectral signal of (a) for fitting to a quadratic polynomial processing method) can be made to be similar in accuracy to processing methods based on lorentzian, gaussian, pseudo-Voigt, and Voigt models, while maintaining fast computations.
3) The invention realizes the rapid measurement of the optical fiber distributed sensing.
Claims (2)
1. A Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting is characterized in that in a measurement signal output by a distributed optical fiber sensor, the Brillouin frequency shift is used as a center to intercept the line width delta v of 1 Brillouin gain spectrumBFitting the signal by a quadratic polynomial processing method to obtain a Brillouin spectrum:
gB(v)＝av2+bv+c
wherein, gB(v) Is the gain of the brillouin scattering spectrum; v is frequency; a, b and c are second-order polynomial coefficients, the numerical values of the coefficients are different along with the change of the Brillouin spectrum, and the Brillouin frequency shift is
2. The method for rapidly and accurately extracting Brillouin frequency shift based on improved quadratic polynomial fitting according to claim 1, wherein the line width Deltav of 1 Brillouin gain spectrum is intercepted by taking the Brillouin frequency shift as a centerBThe brillouin frequency shift is replaced by the frequency corresponding to the brillouin spectral peak.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911309627.4A CN111121836B (en) | 2019-12-18 | 2019-12-18 | Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911309627.4A CN111121836B (en) | 2019-12-18 | 2019-12-18 | Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111121836A true CN111121836A (en) | 2020-05-08 |
| CN111121836B CN111121836B (en) | 2022-02-08 |
Family
ID=70499690
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911309627.4A Active CN111121836B (en) | 2019-12-18 | 2019-12-18 | Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111121836B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111721222A (en) * | 2020-05-19 | 2020-09-29 | 浙江工业大学 | Brillouin frequency shift extraction method based on proximity algorithm |
| CN113566731A (en) * | 2021-08-24 | 2021-10-29 | 广东电网有限责任公司 | Strain calculation method, device, equipment and medium for optical fiber composite overhead line |
| CN113639775A (en) * | 2021-08-11 | 2021-11-12 | 武汉钧恒科技有限公司 | Method and device for frequency shift extraction based on Brillouin optical time domain reflectometer |
| CN113670219A (en) * | 2021-08-16 | 2021-11-19 | 广东电网有限责任公司 | Strain demodulation method, device, equipment and medium for optical fiber line |
| CN113819931A (en) * | 2021-09-28 | 2021-12-21 | 北京卫星环境工程研究所 | BOTDR and BOTDA fusion used Brillouin frequency shift extraction method |
| CN115015312A (en) * | 2022-05-07 | 2022-09-06 | 华中科技大学 | ESR characteristic signal extraction method and system |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002250676A (en) * | 2001-02-26 | 2002-09-06 | Nippon Telegr & Teleph Corp <Ntt> | Method for measuring strain in optical fiber, and instrument therefor |
| CN104457807A (en) * | 2014-11-14 | 2015-03-25 | 南京大学 | Brillouin frequency spectrum peak searching method based on incomplete spectra |
| CN104641201A (en) * | 2012-08-10 | 2015-05-20 | 公益财团法人地球环境产业技术研究机构 | Method for measuring volumetric changes in objects |
| CN106546275A (en) * | 2016-10-31 | 2017-03-29 | 华北电力大学（保定） | A kind of preparation method of Brillouin spectrum characteristic initial value |
| CN108931788A (en) * | 2018-05-29 | 2018-12-04 | 武汉镭英科技有限公司 | A kind of atmosphere Rayleigh Brillouin spectrum solution mixing method based on Viogt model |
-
2019
- 2019-12-18 CN CN201911309627.4A patent/CN111121836B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2002250676A (en) * | 2001-02-26 | 2002-09-06 | Nippon Telegr & Teleph Corp <Ntt> | Method for measuring strain in optical fiber, and instrument therefor |
| CN104641201A (en) * | 2012-08-10 | 2015-05-20 | 公益财团法人地球环境产业技术研究机构 | Method for measuring volumetric changes in objects |
| CN104457807A (en) * | 2014-11-14 | 2015-03-25 | 南京大学 | Brillouin frequency spectrum peak searching method based on incomplete spectra |
| CN106546275A (en) * | 2016-10-31 | 2017-03-29 | 华北电力大学（保定） | A kind of preparation method of Brillouin spectrum characteristic initial value |
| CN108931788A (en) * | 2018-05-29 | 2018-12-04 | 武汉镭英科技有限公司 | A kind of atmosphere Rayleigh Brillouin spectrum solution mixing method based on Viogt model |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111721222A (en) * | 2020-05-19 | 2020-09-29 | 浙江工业大学 | Brillouin frequency shift extraction method based on proximity algorithm |
| CN111721222B (en) * | 2020-05-19 | 2021-08-03 | 浙江工业大学 | Brillouin frequency shift extraction method based on proximity algorithm |
| CN113639775A (en) * | 2021-08-11 | 2021-11-12 | 武汉钧恒科技有限公司 | Method and device for frequency shift extraction based on Brillouin optical time domain reflectometer |
| CN113639775B (en) * | 2021-08-11 | 2023-08-29 | 武汉钧恒科技有限公司 | Frequency shift extraction method and device based on Brillouin optical time domain reflectometer |
| CN113670219A (en) * | 2021-08-16 | 2021-11-19 | 广东电网有限责任公司 | Strain demodulation method, device, equipment and medium for optical fiber line |
| CN113566731A (en) * | 2021-08-24 | 2021-10-29 | 广东电网有限责任公司 | Strain calculation method, device, equipment and medium for optical fiber composite overhead line |
| CN113819931A (en) * | 2021-09-28 | 2021-12-21 | 北京卫星环境工程研究所 | BOTDR and BOTDA fusion used Brillouin frequency shift extraction method |
| CN113819931B (en) * | 2021-09-28 | 2023-06-16 | 北京卫星环境工程研究所 | Brillouin frequency shift extraction method for BOTDR and BOTDA fusion |
| CN115015312A (en) * | 2022-05-07 | 2022-09-06 | 华中科技大学 | ESR characteristic signal extraction method and system |
| CN115015312B (en) * | 2022-05-07 | 2023-03-21 | 华中科技大学 | ESR characteristic signal extraction method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111121836B (en) | 2022-02-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111121836B (en) | Brillouin frequency shift rapid and accurate extraction method based on improved quadratic polynomial fitting | |
| CN107014519B (en) | BOTDR temperature, strain fast separating process in a kind of monitoring of smart grid icing | |
| CN110501092B (en) | Temperature extraction method of Brillouin optical fiber sensing system | |
| Lan et al. | Partial discharge location of power cables based on an improved phase difference method | |
| Wang et al. | Robust and fast temperature extraction for Brillouin optical time-domain analyzer by using denoising autoencoder-based deep neural networks | |
| Zhao et al. | A fast and high accurate initial values obtainment method for Brillouin scattering spectrum parameter estimation | |
| Zou et al. | An efficient cross-terms suppression method in time–frequency domain reflectometry for cable defect localization | |
| CN102818630B (en) | Spectrum calibration method of interference type imaging spectrometer | |
| Zhao et al. | An accurate and rapid method for extracting parameters from multi-peak Brillouin scattering spectra | |
| CN113155170A (en) | Brillouin frequency shift error estimation method | |
| Britton Jr et al. | Johnson noise thermometry for advanced small modular reactors | |
| CN112213560A (en) | High-precision power grid broadband signal measuring method based on Z-ADALINE | |
| CN115856549A (en) | Transformer partial discharge detection positioning method, device and system | |
| CN112989966B (en) | Improved analog circuit signal noise reduction method | |
| CN108845330B (en) | Laser ranging method based on loop frequency measurement | |
| Zhang et al. | Propagation coefficient spectrum based locating method for cable insulation degradation | |
| Lijuan et al. | Influence of optimization model on parameter extraction in Lorentzian Brillouin scattering spectrum | |
| CN113162002A (en) | Direct current traveling wave protection method and system considering broadband measurement link | |
| Zhou et al. | Study on cable fault location based on fast identification of starting point of reflection wave | |
| Wang et al. | Fast peak searching method for Brillouin gain spectrum using positive-slope inflection point | |
| Sun et al. | Gas Insulated Switchgear Partial Discharge Analysis Based on Optical Detection and Positioning | |
| Guo et al. | Temperature Error Compensation Method for FBG Sensor in Electric Power System Based on Fireworks Algorithm | |
| Wang et al. | Simultaneous temperature and strain measurement using deep neural networks for BOTDA sensing system | |
| CN116793533A (en) | BOTDR temperature error estimation method | |
| Song et al. | SPECTRAL CORRECTION METHOD BASED ON IMPROVED FLAT-TOP CONVOLUTED WINDOW FOR PARAMETER ESTIMATION OF POWER HARMONIC. |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |