Frequency spectrum peak searching method for BOTDA (Brillouin Optical Time Domain Analysis) system

Abstract

The invention discloses a spectrum peak searching method for a BOTDA (Brillouin Optical Time Domain Analysis) system. The method comprises the following steps: S1, acquiring Brillouin spectrum data of each point of an optical fiber; s2, performing normalization and noise reduction processing; s3, fitting by using a single-peak Lorentz function to obtain fitting parameters; s4, setting a threshold value, and comparing the spectral width value with the threshold value to judge a single peak or double peaks; s5, calculating a peak value: performing single-peak fitting when the peak value is single, determining a main peak and a side peak when the peak value is double, introducing attenuation to the side peak and reconstructing a signal, and solving a double-peak peak value; and S6, outputting a peak spectrum value. The method can improve the peak searching speed and the measurement precision, reduces the overlapping of sensing information, is suitable for a long-distance system, and improves the sensing performance. The curve fitting error is reduced through normalization and noise reduction, the data comparability is ensured, the fitting calculation amount is simplified, and the multi-peak problem can be effectively solved by judging the double-peak spectrum; the fitting precision and the accuracy of center frequency solution are improved through double-peak fitting, and the calculation amount and the hardware burden are reduced by combining a single-peak and double-peak fitting scheme.

Description

Technical Field

[0001] The present invention belongs to the field of optical fiber sensing, and particularly relates to a method for peak searching of the gain spectrum of a Brillouin distributed optical fiber sensing system. Background Art

[0002] The Brillouin optical time domain reflectometry analyzer (BOTDA) sensing system is a distributed optical fiber sensing technology based on the Brillouin scattering effect in optical fibers. Due to its advantages such as long transmission distance and high measurement accuracy, it is widely used in fields such as the health of building structures, pipeline monitoring and early warning, and temperature and strain monitoring of power cables. Its principle is to collect the Brillouin frequency shift information at various positions of the optical fiber link, and the changes in temperature and strain at each position in the optical fiber are linearly related to the Brillouin frequency shift. Therefore, distributed temperature and strain dual-parameter sensing of optical fibers can be realized.

[0003] However, due to reasons such as optical fiber production and processing or laying, the strain distribution of the optical fiber is uneven, and thus multiple peaks will appear in the measured Brillouin gain spectrum; when the temperature or stress in the optical fiber changes, multiple peaks also occur in the spectrum at the transition position of temperature rise and deformation; the existence of multiple peaks in the gain spectrum will reduce the spatial resolution and measurement accuracy of the system, and increase the misjudgment rate. In addition, for a long-distance BOTDA system, the non-local effect will cause a large difference in the excitation of different frequencies, which will further lead to a decline in the sensing performance of the system.

[0004] After retrieval, a Chinese invention application document with the existing publication number CN102445285B discloses a peak searching method for a BOTDR system, which relates to a Brillouin distributed optical fiber sensing system, "including the following steps: (1) collecting the scattered point data of the Brillouin scattering spectrum; (2) performing data screening to narrow the data range for peak searching; (3) calculating the centroid frequency of the Brillouin scattering spectrum; (4) calculating the Brillouin frequency shift; (5) demodulating the temperature or strain information around the optical fiber according to the change amount of the centroid frequency." However, this method still has the following problems that have not been solved:

[0005] First of all, this method determines the boundary of the Brillouin scattering spectrum through data screening and slope calculation. However, since this method mainly depends on the starting and ending points of the data rather than the complete characteristics of the spectrum, it cannot effectively identify and process the problem of multiple peaks in the measured Brillouin gain spectrum caused by reasons such as optical fiber production and processing or laying, and uneven strain distribution of the optical fiber. Moreover, since this method mainly depends on the starting and ending points of the data to determine the Brillouin scattering spectrum, it may not be able to comprehensively capture the detailed situation of the strain distribution, and thus cannot accurately reflect the uneven strain distribution, especially in areas with complex strain changes.

[0006] Meanwhile, when the temperature or stress in the optical fiber changes, the spectrum at the temperature rise and deformation transition positions also exhibits multiple peaks. In the steps of this method, since the temperature and strain information is demodulated by calculating the centroid frequency and Brillouin frequency shift, and the environment with rapid temperature and strain changes requires a very fast frequency sweeping process to ensure measurement accuracy, this method may not be able to respond in time, resulting in the inability to accurately process the temperature and strain changes in the case of multiple peaks. In addition, this method does not mention how to handle the influence brought by the non-local effect, and the non-local effect will cause a large difference in the excitation of different frequencies, further leading to a decline in the sensing performance of the system. Therefore, this method may have limitations in this regard. Summary of the Invention

[0007] To solve the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a spectrum peak searching method for a BOTDA system, that is, a gain spectrum peak searching method for a Brillouin distributed optical fiber sensing system. The present invention can improve the peak searching speed and measurement accuracy of the Brillouin distributed optical fiber sensing system through an algorithm, and reduce the overlap of sensing information.

[0008] To achieve the above object and other related objects, the present invention adopts the following technical solutions:

[0009] The present invention provides a spectrum peak searching method for a BOTDA system, including the following steps:

[0010] S1. Collect the Brillouin scatter data of each position point of the optical fiber to obtain the Brillouin spectrum I(f, d), where f is the center frequency and d is the position point;

[0011] S2. Perform data preprocessing on the Brillouin spectrum: The data preprocessing includes normalization processing and noise reduction processing;

[0012] S3. Fit the discrete Brillouin spectrum data after data preprocessing with a single-peak Lorentz function to obtain the fitting parameter values;

[0013] S4. Set a threshold, compare the fitting parameter value representing the Brillouin spectral width in the fitting parameters obtained in step S3 with the threshold, and determine whether the Brillouin spectrum is a double-peak spectrum;

[0014] S5. Obtain the peak value:

[0015] When the Brillouin spectrum is a single-peak spectrum, perform single-peak fitting on the discrete Brillouin spectrum data after data preprocessing to obtain the single-peak value;

[0016] When the Brillouin spectrum is a double-peak spectrum, perform double-peak fitting on the discrete Brillouin spectrum data after data preprocessing, determine the main peak f m and the side peak through the change of the contribution ratio, introduce an attenuation amount to the side peak and perform signal reconstruction to restore the position of the main peak and obtain the double-peak value

[0017] S6. Output the peak spectral value obtained in step S5.

[0018] As a preferred technical solution, in step S2, the noise reduction processing includes noise reduction by data point averaging, wavelet noise reduction, or Gaussian filtering.

[0019] As a preferred technical solution, the specific steps of step S3 include:

[0020] Use the simplified formula for fitting, where A 1 ~A 4 are all fitting parameters, and the best fitting parameters (A 1 , A 2 , A 3 , A 4 ) at this point are obtained by fitting according to the LM (Levenberg - Marquardt) algorithm;

[0021] Among them, the fitting parameter A 3 is used to measure the spectral width of the Lorentz curve obtained by fitting the Brillouin spectrum.

[0022] As a preferred technical solution, in step S3, when the fitting result of the Lorentz curve is not good, Gaussian curve fitting or a Lorentz - Gaussian combined model is used, that is, the simplified formula is used for fitting. In the above formula, A 1 ~A 7 are all fitting parameters.

[0023] As a preferred technical solution, in step S4,

[0024] when the fitting parameter representing the Brillouin spectral width among the fitting parameters obtained in step S3 does not exceed the threshold, that is, when the Brillouin spectral width does not exceed the threshold, it is determined that the Brillouin spectrum is a single - peak spectrum;

[0025] when the fitting parameter representing the Brillouin spectral width among the fitting parameters obtained in step S3 exceeds the threshold, that is, when the Brillouin spectral width exceeds the threshold, it is determined that the Brillouin spectrum is a double - peak spectrum; the two peaks in the double - peak spectrum are the result of the superposition of two signals with different center frequencies f 1 and f 2 .

[0026] As a preferred technical solution, in step S5, when fitting the double - peak, by setting the effective interval of parameter iteration, the iteration direction is ensured to be reasonable.

[0027] As a preferred technical solution, in the step S5, when it is determined that the Brillouin spectrum is a double-peak spectrum, according to the result of double-peak fitting, the main peak f is determined by the change of the contribution ratio m and the specific steps of the side peak include:

[0028] First, the two different center frequency signals of the double peak are f 1 and f 2 , and the calculation formula of the contribution ratio of f 1 and f 2 is as follows: Among them, is the peak height of the center frequency f 1 , is the peak height of the center frequency f 2 . In the double-peak interval, the contribution ratios of f 1 and f 2 change dynamically;

[0029] Then, the main peak and the side peak are screened and judged according to the change of the contribution ratio. Among them, the peak value of the main peak increases with the distance, and the peak value of the side peak decreases with the distance.

[0030] As a preferred technical solution, in the step S5, when it is determined that the Brillouin spectrum is a double-peak spectrum, the specific steps of introducing an attenuation amount to the side peak and performing signal reconstruction to restore the accurate position of the main peak include:

[0031] First, introduce to the side peak, where f is the center frequency of the side peak, A is the peak value of the side peak, and B is the spectral width of the side peak;

[0032] Then, reconstruct the Brillouin spectrum to restore the accurate position of the main peak.

[0033] As described above, the present invention has the following beneficial effects:

[0034] (1) A spectrum peak searching method for a BOTDA system according to the present invention improves the peak searching speed and measurement accuracy of the Brillouin distributed optical fiber sensing system, thereby reducing the overlap of sensing information; the peak searching speed and measurement accuracy are effectively improved by the calculation method of the present invention, the overlap of sensing information is reduced, it is applicable to a long-distance BOTDA system, and its sensing performance is improved.

[0035] (2) A spectrum peak searching method for a BOTDA system according to the present invention adopts normalization processing and noise reduction processing, reduces the error of subsequent curve fitting, and ensures the comparability of data; by fitting with a single-peak Lorentz function, the fitting parameter values are obtained, and the fitting iteration calculation amount is simplified; a threshold is set, and by comparing the fitting parameter with the threshold, it is determined whether the Brillouin spectrum is a double-peak spectrum, effectively dealing with the multi-peak problem.

[0036] (3) A spectrum peak searching method for the BOTDA system of the present invention. When the Brillouin spectrum is a double-peak spectrum, the main peak fm and the side peak are determined by the change of the contribution ratio. An attenuation amount is introduced to the side peak and signal reconstruction is performed to restore the accurate position of the main peak and obtain the double-peak value. After double-peak fitting, the fitting accuracy is higher and the central frequency is solved more precisely. At the same time, by adopting a fitting scheme combining single and double peaks, the calculation amount can be effectively reduced and the hardware burden can be alleviated. Brief Description of the Drawings

[0037] Figure 1 It is a schematic flowchart of a spectrum peak searching method for the BOTDA system of the present invention.

[0038] Figure 2 It is a schematic diagram of the data after double-peak fitting of the peak searching algorithm adopted by the present invention. Detailed Embodiments

[0039] In order to better understand the purpose, structure and function of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0040] In the description of the present invention, it should be noted that the positional relationships indicated by terms such as "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. cited in this specification are based on the positional relationships shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific direction, and therefore cannot be construed as a limitation of the present invention.

[0041] Embodiment 1

[0042] As Figure 1 shown, this embodiment provides a spectrum peak searching method for the BOTDA system, that is, a gain spectrum peak searching method for a Brillouin distributed optical fiber sensing system, including the following steps:

[0043] S1. Using the BOTDA system, collect the Brillouin scatter data of each position point of the optical fiber to obtain the Brillouin spectrum I(f, d), where f is the central frequency and d is the position point.

[0044] S2. Perform data preprocessing on the Brillouin spectrum: The data preprocessing includes normalization processing and noise reduction processing to reduce the error of subsequent curve fitting;

[0045] Specifically, in the stage of data preprocessing, the normalization process can eliminate systematic differences under different measurement conditions and ensure the comparability of data. The noise reduction process includes, but is not limited to, noise reduction by data point averaging, wavelet noise reduction, or Gaussian filtering. In this embodiment, noise reduction is performed by averaging data points. In other embodiments, wavelet noise reduction or Gaussian filtering can also be used to reduce the influence of noise on subsequent curve fitting, improve the curve fitting effect, and thus improve the measurement accuracy of the BOTDA system.

[0046] S3. Fit the discrete Brillouin spectrum data after data preprocessing with a single-peak Lorentz function to obtain the fitting parameter values.

[0047] Ideally, the Brillouin spectrum presents a Lorentz function distribution and conforms to the following formula: g 0 is the Brillouin gain coefficient, Δv B is the Brillouin spectral width, v BL is the Brillouin peak frequency. To simplify the fitting iteration calculation amount, a simplified formula is used for fitting, where A 1 ~A 4 are all fitting parameters, and the best fitting parameters (A 1 , A 2 , A 3 , A 4 ) at this point are obtained by fitting according to the LM (Levenberg-Marquardt) algorithm; among them, the fitting parameter A 3 is used to measure the spectral width of the Lorentz curve obtained by fitting the Brillouin spectrum.

[0048] However, in actual engineering implementation, due to factors such as insufficient phonon excitation and limited extinction ratio of the electro-optic modulator, there will be Brillouin spectral broadening, and the Brillouin spectral line shape will transition from a Lorentz curve to a Gaussian curve, resulting in the transformation of the Lorentz line shape to a Gaussian line shape. Therefore, in step S3, when the fitting error of the Lorentz curve exceeds the set deviation value, that is, ∑|x f -x r |>Threshold, where x f is the fitting value, x r is the actual value, and Threshold is the maximum acceptable fitting deviation value set in advance by humans, it is determined that the fitting result of the Lorentz curve is not good. At this time, Gaussian curve fitting or a Lorentz-Gaussian combined model can be tried, that is, a simplified formula is used for fitting. In the above formula, A 1 ~A 7 are all fitting parameters.

[0049] S4. Set a threshold, and compare the fitting parameter A obtained in step S3 with the threshold to determine whether the Brillouin spectrum is a double-peak spectrum. 3 When the fitting parameter A does not exceed the threshold, that is, when the Brillouin spectral width does not exceed the threshold, it is determined that the Brillouin spectrum is a single-peak spectrum.

[0050] When the fitting parameter A 3 exceeds the threshold, that is, when the Brillouin spectral width exceeds the threshold, it is determined that the Brillouin spectrum is a double-peak spectrum. In other words, when the Brillouin spectral width exceeds the threshold, it indicates that there may be an overlap of two peaks in the spectrum at this position. That is, the two peaks in the double-peak spectrum are the result of the superposition of two signals with different central frequencies (f

[0051] When the fitting parameter A 3 This is because in a long-distance BOTDA system, due to the SBS characteristic, the light near the central frequency will be amplified by gain, resulting in different optical excitations at different frequencies at the fiber end. When temperature and strain occur, double peaks may appear. 1 and f 2 ).

[0052] S5. Obtain the peak value.

[0053] When the Brillouin spectrum is a single-peak spectrum, perform single-peak fitting on the discrete Brillouin spectrum data after data preprocessing to obtain the single-peak value.

[0054] When the Brillouin spectrum is a double-peak spectrum, perform double-peak fitting on the discrete Brillouin spectrum data after data preprocessing. According to the result of the double-peak fitting, determine the main peak f m and the side peak. Introduce an attenuation amount to the side peak and perform signal reconstruction to restore the accurate position of the main peak and obtain the double-peak value.

[0055] When performing double-peak fitting, it is necessary to set the effective interval of parameter iteration to ensure a reasonable iteration direction. In this embodiment, the fitting result conforms to the Lorentz curve, so the double-peak Lorentz function is used for double-peak fitting.

[0056] When it is determined that the Brillouin spectrum is a double-peak spectrum, it includes the following specific steps:

[0057] S5-1. Perform double-peak fitting on the discrete Brillouin spectrum data after data preprocessing. According to the result of the double-peak fitting, determine the main peak f m and the side peak, and the main peak is screened by the change of the contribution ratio.

[0058] First, since the two signals with different central frequencies of the double peak are f 1 and f 2 , the calculation formulas for the contribution ratios of f 1 and f 2 are as follows: Among them, is the peak height at the center frequency f 1 , is the peak height at the center frequency f 2 . Within the double-peak interval, the contribution ratios of f 1 and f 2 show dynamic changes.

[0059] Then, the main peak and side peaks are screened and judged according to the changes in the contribution ratios. Among them, the peak value of the main peak increases with the distance, and the peak value of the side peak decreases with the distance. That is, when the Brillouin spectrum is a double-peak spectrum, the one with the higher peak value between the two different center frequencies f at any position point d within the double-peak interval is the main peak fm.

[0060] S5-2. Introduce an attenuation amount to the side peak and perform signal reconstruction.

[0061] First, introduce to the side peak, where f is the center frequency of the side peak, A is the peak value of the side peak, and B is the spectral width of the side peak.

[0062] Then, reconstruct the Brillouin spectrum to restore the accurate position of the main peak.

[0063] Since the main peak reflects the true frequency shift amount and has a strong correlation, it is a correlated peak, while the side peak is an uncorrelated peak. The existence of the uncorrelated peak will cause errors in the actual center frequency value. Therefore, the error can be effectively reduced through value attenuation and signal reconstruction.

[0064] S6. Output the peak spectrum value obtained in step S5.

[0065] As Figure 2 shown, after double-peak fitting, the fitting accuracy is higher and the solution of the center frequency is more accurate. At the same time, by adopting a fitting scheme combining single-peak and double-peak, the calculation amount can be effectively reduced and the hardware burden can be alleviated.

[0066] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any changes or equivalent replacements of these features and embodiments can be made by those skilled in the art without departing from the spirit and scope of the present invention. In addition, under the teaching of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application belong to the scope protected by the present invention.

Claims

1. A spectrum peak finding method for a BOTDA system, characterized in that: The following steps are involved: S1, collect Brillouin scatter data at each position of the optical fiber to obtain the Brillouin spectrum I(f,d), where f is the center frequency and d is the position point; S2. performing data preprocessing on the Brillouin spectrum: the data preprocessing includes normalization processing and noise reduction processing; S3, fitting the discrete Brillouin spectrum data after data preprocessing using a single-peak Lorentz function to obtain fitting parameter values; S4, setting a threshold, comparing the fitting parameter value representing the Brillouin spectrum width in the fitting parameters obtained in step S3 with the threshold, and determining whether the Brillouin spectrum is a double-peak spectrum; S5. Find the peak value: When the Brillouin spectrum is a single-peak spectrum, a single-peak fitting is performed on the discrete Brillouin spectrum data after data preprocessing to obtain the single-peak peak value; When the Brillouin spectrum is a double-peak spectrum, the discrete Brillouin spectrum data after data preprocessing is fitted with a double peak, and the main peak f is determined by the change of the contribution ratio. m and side peaks, introduce attenuation to the side peaks and reconstruct the signal to restore the main peak position and obtain the double peak peak S6. Output the peak spectrum value obtained in step S5.

2. A spectrum peak finding method for a BOTDA system according to claim 1, characterized in that: In the step S2, the noise reduction process includes data point averaging noise reduction, wavelet noise reduction or Gaussian filtering.

3. The spectrum peak finding method for a BOTDA system according to claim 1, characterized in that: The specific steps of step S3 include: Using the simplified formula Fitting is performed, where A1~A4 are all fitting parameters, and the best fitting parameters (A1, A2, A3, A4) at this point are obtained by fitting according to the LM (Levenberg-Marquardt) algorithm; The fitting parameter A3 is used to measure the spectral width of the Lorentz curve obtained by Brillouin spectrum fitting.

4. The spectrum peak finding method for a BOTDA system according to claim 1, characterized in that: In step S3, when the Lorentz curve fitting result is not good, Gaussian curve fitting or Lorentz-Gaussian combination model is adopted, that is, a simplified formula is adopted. Perform fitting, where A1~A7 are all fitting parameters.

5. The spectrum peak finding method for a BOTDA system according to claim 1, characterized in that: In step S4, When the value of the fitting parameter representing the Brillouin spectrum width among the fitting parameters obtained in step S3 does not exceed the threshold value, that is, the Brillouin spectrum width does not exceed the threshold value, the Brillouin spectrum is judged to be a single-peak spectrum; When the fitting parameter value representing the Brillouin spectrum width in the fitting parameters obtained in step S3 exceeds the threshold, that is, the Brillouin spectrum width exceeds the threshold, the Brillouin spectrum is judged to be a double-peak spectrum; the double peaks in the double-peak spectrum are the result of superposition of two different center frequency signals f1 and f2.

6. A spectrum peak finding method for a BOTDA system according to claim 1, characterized in that: In step S5, during bimodal fitting, the parameter iteration valid interval is set to ensure that the iteration direction is reasonable.

7. The spectrum peak finding method for a BOTDA system according to claim 5, characterized in that: In step S5, when the Brillouin spectrum is determined to be a double-peak spectrum, the main peak f is determined by the change of the contribution ratio according to the result of the double-peak fitting. m The specific steps of side peak include: First, the two different center frequency signals of the double peak are f1 and f2 respectively, and the contribution ratio of f1 and f2 is calculated as follows: in, is the peak height of the center frequency f1, is the peak height of the center frequency f2, and the contribution ratio of f1 and f2 changes dynamically within the double-peak interval; Then, the main peak and the side peak are screened and judged according to the change of the contribution ratio, wherein the peak value of the main peak increases with the distance, and the peak value of the side peak decreases with the distance.

8. A spectrum peak finding method for a BOTDA system according to claim 7, characterized in that: In step S5, when it is determined that the Brillouin spectrum is a double-peak spectrum, the specific steps of introducing attenuation to the side peaks and performing signal reconstruction to restore the accurate main peak position include: First, an attenuation is introduced into the side peak Where, f is the center frequency of the side peak, A is the peak value of the side peak, and B is the spectrum width of the side peak; Then, the Brillouin spectrum is reconstructed to restore the accurate main peak position.

Images