JP7533382B2 - Spectral analysis device, spectral analysis method, and spectral analysis program - Google Patents

Description

The present invention relates to a spectrum analysis device, a spectrum analysis method, and a spectrum analysis program.

The spectrum (frequency) of the Brillouin scattered light generated by irradiating light into an optical fiber changes with changes in the temperature and strain of the optical fiber. Optical fiber sensors that utilize this property measure the temperature and strain distribution along the length of the optical fiber by detecting changes in the frequency of the Brillouin scattered light (Brillouin frequency shift (BFS)) along the length of the optical fiber.

Typical examples of such optical fiber sensors include the BOTDR (Brillouin Optical Time Domain Reflectometry) type disclosed in Patent Document 1 below, and the BOTDA (Brillouin Optical Time Domain Analysis) type disclosed in Patent Document 2 below. As shown in Non-Patent Documents 1 and 2 below, it is known that when the optical fiber used for measurement straddles two regions with different temperatures or strains, two peaks (double peaks) appear in the spectrum of the measured Brillouin scattered light.

Patent No. 5168839 Patent No. 2589345

Y. Yu, L. Luo, B. Li, L. Guo, J. Yan, and K. Soga, “Double peak-induced distance error in short-time-Fourier-transform-BOTDR event detection and the recovery method,” Applied Optics Vol.54, No.28, 2015. O. Furukawa, “Prevention of false peak detection of Brillouin gain spectrum by using peak tracking and trend analysis,” Electron. Comm. Japan, Vol.103, 2020.

As shown in the above-mentioned non-patent documents 1 and 2, when two peaks appear in the spectrum of Brillouin scattered light, if the intensity of one of the peaks is significantly higher than the intensity of the other peak, there is almost no chance of the maximum peak being erroneously detected. However, when the intensities of the two peaks are competitive, there is a problem in that, for example, noise inherent to the optical fiber sensor is superimposed on the spectrum of the Brillouin scattered light, causing the maximum peak to be erroneously detected.

The present invention has been made in consideration of the above circumstances, and aims to provide a spectrum analysis device, a spectrum analysis method, and a spectrum analysis program that can prevent erroneous detection of the maximum peak when two peaks appear in the spectrum of Brillouin scattered light.

In order to solve the above problem, a spectrum analysis device (10) according to one aspect of the present invention includes an acquisition unit (11) that acquires spectrum data indicating the spectrum of Brillouin scattered light generated at different positions in the longitudinal direction of an optical fiber (FUT), a separation and extraction unit (13) that separates and extracts a first peak and a second peak for each of the spectrum data, an identification unit (15) that identifies a position on the optical fiber where the intensity of the first peak and the intensity of the second peak are close to each other and where the spectrum data was obtained, a gradient calculation unit (16) that calculates the gradient of the intensity of the first peak and the gradient of the intensity of the second peak at the position identified by the identification unit, and a determination unit (17) that determines which of the first peak and the second peak is appropriate as the maximum peak based on the gradient calculated by the gradient calculation unit.

The spectral analysis device according to one aspect of the present invention further includes a normalization unit (12) that normalizes the intensity of each of the spectral data acquired by the acquisition unit.

In addition, in one aspect of the present invention, the acquisition unit of the spectrum analysis device acquires the spectrum data detected by a BOTDR or BOTDA optical fiber sensor.

In addition, in a spectral analysis device according to one aspect of the present invention, the separation and extraction unit performs curve fitting using a Lorentz distribution function or a Voigt function having an arbitrary half-width for each of the spectral data to separate and extract the first peak and the second peak.

In addition, in a spectral analysis device according to one aspect of the present invention, the slope calculation unit performs linear regression analysis on the first peak at the position identified by the identification unit and on the first peak at positions before and after the position to calculate the slope of the intensity of the first peak, and performs linear regression analysis on the second peak at the position identified by the identification unit and on the second peak at positions before and after the position to calculate the slope of the intensity of the second peak.

The spectrum analysis device according to one aspect of the present invention also includes a peak frequency output unit (18) that outputs the frequency of the peak that is determined by the determination unit to be appropriate as the maximum peak.

The spectrum analysis device according to one aspect of the present invention also includes a fiber characteristic output unit (19) that outputs the characteristics of the optical fiber determined from the frequency output from the peak frequency output unit.

The spectral analysis method according to one aspect of the present invention includes an acquisition step (S11) of acquiring spectral data showing the spectrum of Brillouin scattered light generated at different positions in the longitudinal direction of an optical fiber (FUT), a separation and extraction step (S13) of separating and extracting a first peak and a second peak for each of the spectral data, a specification step (S15) of identifying a position on the optical fiber where the intensity of the first peak and the intensity of the second peak are close to each other and where the spectral data was obtained, a slope calculation step (S16) of calculating the slope of the intensity of the first peak and the slope of the intensity of the second peak at the position identified by the specification step, and a determination step (S17) of determining which of the first peak and the second peak is appropriate as the maximum peak based on the slope calculated in the slope calculation step.

The spectral analysis program according to one aspect of the present invention causes a computer to function as an acquisition means (11) for acquiring spectral data showing the spectrum of Brillouin scattered light generated at different positions in the length direction of an optical fiber (FUT), a separation and extraction means (13) for separating and extracting a first peak and a second peak for each of the spectral data, a determination means (15) for determining a position on the optical fiber where the intensity of the first peak and the intensity of the second peak are close to each other and where the spectral data was obtained, a slope calculation means (16) for calculating the slope of the intensity of the first peak and the slope of the intensity of the second peak at the position determined by the determination means, and a determination means (17) for determining which of the first peak and the second peak is appropriate as the maximum peak based on the slope calculated by the slope calculation means.

The present invention has the effect of preventing erroneous detection of the maximum peak when two peaks appear in the spectrum of Brillouin scattered light.

1 is a block diagram showing a configuration of a main part of a spectrum analysis device according to an embodiment of the present invention; FIG. 4 is a diagram for explaining a process performed in a peak separation and extraction section in one embodiment of the present invention. FIG. 11 is a diagram for explaining another process performed in the peak separation and extraction section in the embodiment of the present invention. 5 is a diagram for explaining a process performed in a peak intensity gradient calculation unit in the embodiment of the present invention. FIG. FIG. 11 is a diagram for explaining a peak information correction process performed in one embodiment of the present invention. 2 is a flow chart illustrating a spectral analysis method according to an embodiment of the present invention.

The following describes in detail a spectrum analysis device, a spectrum analysis method, and a spectrum analysis program according to an embodiment of the present invention with reference to the drawings. First, an overview of the embodiment of the present invention will be described, followed by a detailed description of the embodiment of the present invention.

〔overview〕
The embodiment of the present invention makes it possible to prevent erroneous detection of the maximum peak when two peaks appear in the spectrum of Brillouin scattered light. Specifically, when the intensities of two peaks appearing in the spectrum of Brillouin scattered light are competitive, for example, even if noise inherent to the optical fiber sensor is superimposed on the spectrum of Brillouin scattered light, it makes it possible to prevent erroneous detection of the maximum peak.

The above-mentioned non-patent document 1 discusses that when two peaks appear in the spectrum of Brillouin scattered light, an error occurs in the detection distance of the Brillouin frequency shift. However, there is no discussion of the false detection of the maximum peak that can occur when the intensities of the two peaks are competitive. The frequency of occurrence of this problem of false detection of the maximum peak can be improved by reducing the noise of the optical fiber sensor, but this alone cannot completely solve the problem.

The above-mentioned non-patent document 2 discloses a peak tracking trend analysis method that can solve the problem when two peaks appear in the spectrum of Brillouin scattered light. This peak tracking trend analysis method has a first step (peak tracking) and a second step (trend analysis).

In the first step (peak tracking), the maximum peak and the next largest peak are found in the spectrum of the Brillouin scattered light (hereinafter sometimes referred to as the "Brillouin scattered spectrum"), and peak extraction is performed assuming a Lorentzian distribution. Next, group numbers are assigned to classify the two peaks according to their frequencies. The peak frequency and spectral intensity are then stored in the group, and peaks that fall within a specified frequency range from the stored peak frequency are treated as belonging to the same group as the group number. Then, if the peak frequency changes along the length of the optical fiber and they are in the same group, the peak frequency stored in the group is updated, and the maximum peak and the next largest peak are tracked and recorded by group number.

In the second step (trend analysis), first, the number of consecutive occurrences of the same recorded group number is counted. Next, if the spectral intensity rises/falls in the count, the number of occurrences is increased/decreased, and the increasing/decreasing count value is used to analyze the rising/falling trend of the spectral intensity. When the peak frequencies contained in two groups transition from one to the other, it is natural that the spectral intensity at the frequency that becomes the new maximum peak will trend upwards. For this reason, a transition from an upward trending peak to a downward trending peak should be interpreted as a false positive, and the transition is deemed invalid.

In this type of peak tracking trend analysis method, the first step (peak tracking) is performed on all measurement data along the length of the optical fiber. If the intensity of the largest peak is not competitive with the intensity of the next largest peak, there is no possibility of a false positive and the process is wasted. In addition, the predetermined frequency range used for grouping in the first step (peak tracking) is set as a threshold value based on the intrinsic full width at half maximum spectrum of Brillouin scattering, but there is a constraint that false positives are not judged within that range.

In an embodiment of the present invention, first, spectral data showing the spectrum of Brillouin scattered light generated at different positions in the length direction of the optical fiber is obtained, and a first peak and a second peak are separated and extracted for each of the spectral data. Next, a position on the optical fiber where the spectral data was obtained where the intensity of the first peak and the intensity of the second peak are close to each other is identified, and the gradient of the intensity of the first peak and the gradient of the intensity of the second peak at the identified position are calculated. Then, based on the calculated gradient, it is determined which of the first peak and the second peak is appropriate as the maximum peak. This makes it possible to prevent erroneous detection of the maximum peak when two peaks appear in the spectrum of the Brillouin scattered light.

Furthermore, in an embodiment of the present invention, unnecessary processing performed in the peak tracking trend analysis method disclosed in the above-mentioned non-patent document 2 is eliminated, and the restriction on the frequency range in which false detection judgment is not performed is eliminated, making it possible to prevent false detection of the maximum peak with a simpler method. Specifically, by focusing only on the trend analysis of the Brillouin scattering spectrum intensity and using the gradient of the intensity of each peak when the intensities of two peaks are close to each other, false detection of the maximum peak can be prevented.

〔detail〕
Spectral Analysis Device
Fig. 1 is a block diagram showing the main components of a spectral analysis device according to an embodiment of the present invention. As shown in Fig. 1, the spectral analysis device 10 of this embodiment includes a spectral data acquisition unit 11 (acquisition unit, acquisition means), a spectral intensity normalization unit 12 (normalization unit), a peak separation and extraction unit 13 (separation and extraction unit, separation and extraction means), a peak information storage unit 14, a peak intensity proximity identification unit 15 (identification unit, identification means), a peak intensity slope calculation unit 16 (slope calculation unit, slope calculation means), a peak determination unit 17 (determination unit, determination means), a peak frequency output unit 18, and a fiber characteristic output unit 19.

The above-described spectrum analyzer 10, together with the optical fiber sensor body 20, constitutes the optical fiber sensor system 1. The optical fiber sensor system 1 is a system in which the spectrum analyzer 10 acquires and analyzes the spectrum data detected by the optical fiber sensor body 20 to determine the characteristics (distribution of temperature and strain) along the length of the optical fiber FUT.

The optical fiber sensor body 20 may be of the BOTDR (Brillouin Optical Time Domain Reflectometry) type or the BOTDA (Brillouin Optical Time Domain Analysis) type.

The above BOTDR method involves injecting pulsed light into one end of the optical fiber FUT. The characteristics of the optical fiber FUT are then determined by sequentially detecting the backscattered Brillouin light emitted from the same end of the optical fiber FUT.

In the BOTDA method, a pulsed light (pump light) with a variable frequency is input from one end of the optical fiber FUT, and a probe light, which is a continuous light, is input from the other end of the optical fiber FUT. The characteristics of the optical fiber FUT are obtained by sequentially measuring the changing components of the strong scattered light generated by the stimulated Brillouin scattering phenomenon.

The spectral data acquisition unit 11 acquires the spectral data output from the optical fiber sensor body 20. Here, the spectral data is data that indicates the spectrum of the Brillouin scattered light generated within the optical fiber FUT. The optical fiber sensor body 20 typically performs measurements at multiple measurement points set along the length of the optical fiber FUT. For this reason, the spectral data acquisition unit 11 acquires spectral data that indicates the spectrum of the Brillouin scattered light generated at each measurement point.

The spectral data acquired by the spectral data acquisition unit 11 is three-dimensional data that combines information on the position (distance) in the longitudinal direction of the optical fiber FUT, information on the frequency axis of the Brillouin scattering spectrum, and information on the intensity of the Brillouin scattering spectrum. If a large amount of noise is superimposed on the spectral data acquired by the spectral data acquisition unit 11, the noise may be reduced by using low-pass filter processing.

The spectral intensity normalization unit 12 normalizes the intensity of each piece of spectral data acquired by the spectral data acquisition unit. For example, the spectral intensity normalization unit 12 normalizes the intensity of each piece of spectral data so that the area of the Brillouin scattering spectrum becomes a predetermined value (e.g., "1"). Such normalization is performed because the light intensity of the Brillouin scattering spectrum obtained from the optical fiber FUT varies due to factors such as the material, temperature characteristics, and transmission distance of the optical fiber FUT.

For example, if the optical fiber FUT is a fusion spliced or connectorized combination of multiple optical fibers, the light intensity of the Brillouin scattering spectrum may differ depending on the material of each optical fiber. Even if the optical fiber FUT is a single optical fiber with no breaks in its length, it is known that the light intensity of the Brillouin scattering spectrum obtained from a location with a high ambient temperature is greater than the light intensity of the Brillouin scattering spectrum obtained from a location with a low ambient temperature. Furthermore, because the optical fiber FUT has propagation loss, the light intensity of the Brillouin scattering spectrum decreases as the propagation distance increases. The above normalization is performed to eliminate changes in light intensity caused by such factors.

The peak separation extraction unit 13 separates and extracts at least two peaks, the maximum peak and the second largest peak, for each of the spectral data normalized by the spectral intensity normalization unit 12. It is known that the peaks that make up the Brillouin scattering spectrum are originally Lorentzian distributions. For this reason, the peak separation extraction unit 13 can also separate and extract the above two peaks from a spectrum represented by a linear sum of a Lorentzian distribution function having an arbitrary half-width, or a Voigt function in which a Gaussian distribution with a random component is added to a Lorentzian distribution having an arbitrary half-width.

Here, if only the above two peaks are used, the accuracy of the Brillouin frequency shift (BFS) extraction may deteriorate. For this reason, it is possible to improve accuracy by extracting three or more peaks separately or by applying the Shirly method, which removes the slope of the baseline. Methods for finding peaks include a local search method that simply finds the maximum value, a method that finds the point where the first derivative crosses zero, a method that finds a point that is several times the standard deviation away from the average value, and a method that uses curve fitting.

Figure 2 is a diagram for explaining the processing performed in the peak separation and extraction unit in one embodiment of the present invention. The graph shown in Figure 2 shows the Brillouin scattering spectrum, with frequency on the horizontal axis and light intensity on the vertical axis. In Figure 2(a), spectrum BS1 plotted with circles is data obtained by sampling and normalizing the Brillouin scattering spectrum at a specified frequency interval (spectral data whose intensity has been normalized by the spectrum intensity normalization unit 12).

The peak separation extraction unit 13 searches for a peak in the spectrum BS1 using a local search method. The spectrum BS1 illustrated in FIG. 2A contains two peaks whose peak frequencies are close to each other, but the local search method can only find one peak P _1m (frequency f ₁ , light intensity I _1m ). If only one peak can be found, the peak separation extraction unit 13 subtracts the Lorentz distribution D1 (frequency f ₁ , light intensity I _1m ) shown by the dashed line from the spectrum BS1 to obtain the spectrum BS2 shown in FIG. 2B. Then, the peak separation extraction unit 13 searches for a peak in the spectrum BS2 using a local search method. This makes it possible to find the peak P _2m (frequency f ₂ , light intensity I _2m ) that is the next largest after the peak P _1m .

By the above process, the maximum peak, peak _P1m , and the next largest peak, peak _P2m , are separated and extracted. However, if the frequency interval of the spectrum data is not sufficiently precise, the frequencies and optical intensities of peaks _P1m and _P2m separated and extracted by the above process are not accurate, and so these are used to provide initial values for curve fitting, which will be described next.

Figure 3 is a diagram for explaining another process performed in the peak separation and extraction unit in one embodiment of the present invention. The graph shown in Figure 3, like the graph shown in Figure 2, shows the Brillouin scattering spectrum, with frequency on the horizontal axis and light intensity on the vertical axis. In Figure 3(a), the spectrum BS plotted with circles is data obtained by sampling and normalizing the Brillouin scattering spectrum at a specified frequency interval (spectral data whose intensity has been normalized by the spectrum intensity normalization unit 12).

The peak separation and extraction unit 13 performs curve fitting of a quadratic function for several samples on the left and right of the two peaks _P1m and _P2m shown in Fig. 2. For example, a quadratic curve C1 is obtained by performing curve fitting on samples in a region R1 including several samples on the left and right of the peak _P1m . A quadratic curve C2 is obtained by performing curve fitting on samples in a region R2 including several samples on the left and right of the peak _P2m . From the coefficients of the quadratic curves C1 and C2, a first peak (peak on the high frequency side) having a frequency _f1 and a light intensity _I1 is obtained near the maximum value peak _P1m , and a second peak (peak on the low frequency side) having a frequency _f2 and a light intensity _I2 is obtained near the next largest peak _P2m .

The peak separation and extraction unit 13 can also perform curve fitting using a Lorentz distribution function having an arbitrary half-width, or a Voigt function obtained by adding a Gaussian distribution with a random component to a Lorentz distribution having an arbitrary half-width, instead of a quadratic function. For example, as shown in FIG. 3B, curves L1 and L2 are obtained by curve fitting a curve defined by the linear sum of two functions to a sample of the spectrum BS. As a result, similar to the example shown in FIG. 3A, a first peak on the high frequency side having a frequency f ₁ and a light intensity I ₁ is obtained near the maximum value peak P _1m , and a second peak on the low frequency side having a frequency f ₂ and a light intensity I ₂ is obtained near the next largest peak P _2m .

The frequency and light intensity of the peaks (first and second peaks) obtained in the explanation of FIG. 2 may be used as the initial values for the curve fitting explained in FIG. 3(b). The light intensity may also be adjusted using the matrix calculation shown in the aforementioned non-patent document 2. Furthermore, the peak separation and extraction method performed by the peak separation and extraction unit 13 may be another peak separation and extraction method generally known as a spectral analysis method in spectroscopy.

The peak information storage unit 14 stores the peak information (frequency and intensity) separated and extracted by the peak separation and extraction unit 13 together with the position (distance from one end of the optical fiber FUT). Information storage is achieved by a recording medium that can be read by a computer or electronic circuit. Recording media include, for example, magnetic recording media (hard disk drives, magneto-optical disks, etc.) and semiconductor memories (flash memories, random access memories (RAMs)).

For each piece of spectral data, if the intensity of the first peak and the intensity of the second peak are close to each other, the peak intensity proximity identification unit 15 identifies the position on the optical fiber FUT where the spectral data was obtained (the distance from one end of the optical fiber FUT). The peak intensity proximity identification unit 15 determines whether the intensity of the first peak and the intensity of the second peak are close to each other by whether the difference between the intensity of the first peak and the intensity of the second peak is equal to or less than a predefined threshold value.

For example, when the light intensity I is expressed in decibel units such as 5 log ₁₀ (I), if the difference between the intensity of the first peak and the intensity of the second peak is within a threshold value of 0.1 dB, the peak intensity proximity identifying unit 15 determines that the intensity of the first peak and the intensity of the second peak are close to each other. This threshold value can be changed, for example, depending on the level of noise in the measuring device of the optical fiber sensor main body 20. The number of spectrum data (the number of identified positions on the optical fiber FUT) for which the peak intensity proximity identifying unit 15 has determined that the intensity of the first peak and the intensity of the second peak are close to each other is denoted by u.

The peak intensity slope calculation unit 16 calculates the slope of the peak intensity in the length direction of the optical fiber FUT at each of the u positions identified as having close peak intensities. The peak intensity slope calculation unit 16 calculates the slope of the peak intensity using the peak intensity at the position where the slope of the peak intensity is to be calculated, as well as the peak intensity at positions before and after that position. The peak intensity slope calculation unit 16 calculates the slope of the peak intensity separately for the low frequency side and the high frequency side. That is, the peak intensity slope calculation unit 16 uses only the peak intensity on the low frequency side (the intensity of the second peak) when calculating the slope of the peak intensity on the low frequency side, and uses only the peak intensity on the high frequency side (the intensity of the first peak) when calculating the slope of the peak intensity on the high frequency side.

Figure 4 is a diagram for explaining the processing performed by the peak intensity slope calculation unit in one embodiment of the present invention. The graph shown in Figure 4 is a graph showing the intensity change of the first peak and the second peak in the longitudinal direction of the optical fiber FUT, with the horizontal axis representing the position on the optical fiber FUT and the vertical axis representing the light intensity. Note that the origin position of the horizontal axis of the graph shown in Figure 4 indicates the position of one end of the optical fiber FUT.

As shown in Fig. 4, for example, assume that the difference between the intensity of the first peak and the intensity of the second peak separated and extracted from the kth spectrum data obtained at a position _dk (k is an integer) on the optical fiber FUT is equal to or less than a predefined threshold value. In Fig. 4, the first peak having a high frequency (peak having a frequency _fH ) is shown by a circle plot, and the second peak having a low frequency (peak having a frequency _fL ) is shown by a diamond plot.

In the example shown in FIG. 4, at the origin position (position 0), the light intensity of the first peak (peak with a frequency of _fH ) with a high frequency is higher than the light intensity of the second peak (peak with a frequency of _fL ) with a low frequency. As the distance from the origin position increases, the intensity of the first peak (peak with a frequency of _fH ) with a high frequency decreases, and the intensity of the second peak (peak with a frequency of _fL ) with a low frequency increases. Then, at the position _dk-2 on the optical fiber FUT, the magnitude relationship between the intensity of the first peak and the intensity of the second peak is reversed. As a result, the second peak (peak with a frequency of _fL ) with a low frequency becomes the new maximum peak, and the frequency of the maximum peak transitions from the high frequency side (frequency _f1 ) to the low frequency side (frequency _f2 ). In such a frequency transition, there is a possibility that a double peak may be erroneously detected.

After the above frequency transition, the correct result is that the second peak with a lower frequency (the peak with a frequency of _fL ) becomes the maximum peak. However, in the example shown in Fig. 4, at the position _dk on the optical fiber FUT, the magnitude relationship between the intensities of the first peak and the second peak is reversed again due to the influence of the superimposed noise, and the first peak with a higher frequency (the peak with a frequency of _fH ) becomes the maximum peak.

The peak intensity slope calculation unit 16 calculates the slope of the peak intensity in order to determine whether such a reversal of the peak intensity is caused by noise. The peak intensity slope calculation unit 16 calculates the slope of the peak intensity by linear regression using the peak intensity at a plurality of positions (for example, five positions from positions d _k-2 to d _{k +2} ) before and after the position d _k on the optical fiber FUT. Note that other calculation methods may include the least squares method or curve fitting of a monotonic function. The slope of the peak intensity on the high frequency side calculated by the peak intensity slope calculation unit 16 is S _H , and the slope of the peak intensity on the low frequency side is S _L.

The peak determination unit 17 uses the slope of the peak intensity calculated by the peak intensity slope calculation unit 16 to determine whether the first peak or the second peak is appropriate as the maximum peak. Specifically, the peak determination unit 17 determines whether the frequency transition of the maximum peak at the position identified by the peak intensity proximity identification unit 15 is caused by the influence of noise or the like, thereby determining whether the first peak or the second peak is appropriate as the maximum peak.

Here, since the Brillouin scattering spectrum obtained from the correct peak position has a high intensity, it is natural that the transition to the correct peak is accompanied by a gradual increase in the spectrum intensity. At the position d _k shown in FIG. 4, the slope S _H of the peak intensity on the high frequency side, which has become a new maximum peak due to the frequency transition, is negative (S _H <0), and the slope S _L of the peak intensity on the low frequency side is positive (S _L >0). Therefore, the peak determination unit 17 determines that the transition from the low frequency side to the high frequency side at the position d _k is caused by the influence of noise or the like (false detection), and that the second peak with a lower frequency (peak with a frequency f _L ) is appropriate as the maximum peak at this position. The peak information stored in the peak information storage unit 14 is corrected by the peak determination unit 17 based on the determination result of the peak determination unit 17.

More specifically, the peak determination unit 17 obtains the difference (S _H -S _L ) between the slope S _H of the peak intensity on the high frequency side and the slope S _L of the peak intensity on the low frequency side, and performs the above-mentioned determination in accordance with the contents of the following Table 1. Note that if S _H -S _L < 0, the second peak with a lower frequency (the peak with a frequency f _L ) is the maximum peak, and if S _H -S _L > 0, the first peak with a higher frequency (the peak with a frequency f _H ) is the maximum peak.

Here, when the difference in the slope of the peak intensities is 0 (when S _H - _{S L} = 0), the frequency transition of the maximum peak may be determined to be an erroneous detection, or may be determined to be a correct transition. Alternatively, when S _H - _{S L} = 0, the number of points used in the peak intensity slope calculation unit 16 may be increased (for example, from five points including position d _k to seven points) and the slope may be recalculated, and then S _H - S _L may be calculated again to make a judgment.

The peak determination unit 17 determines which of the first peak and the second peak is appropriate as the maximum peak for each of the u positions identified by the peak intensity proximity identification unit 15 as having close peak intensities. However, depending on the determination result in the processing process, the number of processing times may be less than that for the u positions. For example, when the frequency of the maximum peak transitions from the high frequency side (frequency f ₁ ) to the low frequency side (frequency f ₂ ) and then transitions back to the high frequency side (frequency f ₁ ), if it is determined that the first transition is an erroneous detection, there is no need to determine the next transition.

Figure 5 is a diagram for explaining the peak information correction process performed in one embodiment of the present invention. The graph shown in Figure 5 is a graph showing the Brillouin frequency shift (BFS) in the longitudinal direction of the optical fiber FUT, with the horizontal axis representing the position on the optical fiber FUT and the vertical axis representing the Brillouin frequency shift (BFS). Note that the origin position of the horizontal axis of the graph shown in Figure 5 indicates the position of one end of the optical fiber FUT.

For ease of explanation, it is assumed here that the intensity changes of the first peak (peak with frequency _fH ) and the second peak (peak with frequency _fL ) in the longitudinal direction of the optical fiber FUT are the same as those shown in Fig. 4. In other words, it is assumed that the magnitude relationship between the intensity of the first peak and the intensity of the second peak is reversed at position _dk-2 on the optical fiber FUT, and is reversed again at position _dk on the optical fiber FUT.

When such an intensity change occurs, as shown in Fig. 5(a), the Brillouin frequency shift transitions from frequency _fH to frequency _fL at position _dk-2 , and transitions from frequency _fL to frequency _fH at position _dk . When the peak determination unit 17 determines that the transition at position _dk (the transition from frequency _fL to frequency _fH ) is an erroneous detection, it corrects the peak information stored in the peak information storage unit 14 so that the Brillouin frequency shift at position _dk becomes frequency _fL , as shown in Fig. 5(b).

Here, the interval between position d _k and the adjacent position (position d _k-1 or position d _k+1 ) depends on the setting when performing measurement with the optical fiber sensor body 20, but is generally called the sampling interval or sampling resolution. The sampling interval of the BOTDR is about 1 m. This sampling interval needs to be sufficiently narrower (higher sampling resolution) than the area where temperature and strain change.

For example, suppose that the optical fiber sensor body 20 is a BOTDR with a sampling interval of 1 m, and the optical fiber sensor system 1 measures the strain occurring in a structure by measuring the strain applied to the optical fiber FUT that is laid throughout the structure. If the structure to be measured is one in which the strain changes significantly in units of 10 cm, the peak intensity slope calculation unit 16 cannot properly obtain the slope of the peak intensity. For this reason, the peak determination unit 17 cannot determine whether the first peak or the second peak is appropriate as the maximum peak.

In contrast, when the structure to be measured is a structure (e.g., a pipeline) in which the distortion changes gradually over several times the sampling interval (e.g., 10 m), the peak intensity gradient calculation unit 16 can appropriately obtain the gradient of the peak intensity. Therefore, the peak determination unit 17 can determine which of the first peak and the second peak is appropriate as the maximum peak. Note that a prerequisite for using the optical fiber sensor system 1 is that the sampling interval of the optical fiber sensor body 20 is shorter than the spatial resolution (theoretical limit for identifying peaks) of the optical fiber sensor body 20.

Also, for example, assume that the optical fiber sensor body 20 is a BOTDR with a sampling interval of 50 cm and a spatial resolution of 1 m, and the optical fiber sensor system 1 measures temperature changes rather than strain. In such an optical fiber sensor system 1, if the sampling interval and spatial resolution are appropriate, it is possible to determine which of the first peak and the second peak is appropriate as the maximum peak.

Here, a spatial resolution of 1 m means that the temperature can be accurately measured within a range of position d _k ±50 cm. Furthermore, a sampling interval of 50 cm means that the interval between positions d _k-1 and d _k is 50 cm. In a situation where the temperatures at adjacent positions every 50 cm are significantly different and discontinuous (e.g., 0° C. and 200° C.), it is possible to determine whether the first peak or the second peak is the maximum peak by appropriately obtaining the slope of the spectral intensity.

The peak frequency output unit 18 outputs the peak frequency of the maximum peak after it is determined by the peak determination unit 17 to be a false detection and corrected to the correct frequency. The output of the peak frequency output unit 18 may include the Brillouin frequency shift not only at the position u where the peak intensity proximity determination unit 15 has determined that the peak intensities are close, but also at all or some of the positions distributed along the length of the optical fiber FUT.

The fiber characteristic output unit 19 outputs the characteristics (strain and temperature) of the optical fiber FUT obtained from the Brillouin frequency shift output from the peak frequency output unit 18. Specifically, the fiber characteristic output unit 19 outputs the characteristics (strain and temperature) of the optical fiber FUT obtained by multiplying the above-mentioned Brillouin frequency shift by a predetermined conversion coefficient corresponding to the optical fiber FUT.

The peak frequency output unit 18 and the fiber characteristic output unit 19 may be equipped with a display that outputs the peak frequency (Brillouin frequency shift) and the characteristics (strain and temperature) of the optical fiber FUT. Examples of the display include a cathode ray tube (CRT), a liquid crystal display (LCD), an organic electroluminescent (EL) display, a plasma display (PDP), and projection type technology.

Spectral analysis method
6 is a flowchart showing a spectrum analysis method according to an embodiment of the present invention. The process of the flowchart shown in FIG. 6 is started, for example, when an instruction or operation is received from a user of the spectrum analysis apparatus 10.

When the process of the flowchart shown in FIG. 6 is started, first, the spectral data acquisition unit 11 acquires the spectral data output from the optical fiber sensor body 20 (step S11: acquisition step). The number of times that the spectral data is acquired may be one or more. In practice, the spectral data may be acquired multiple times (e.g., 100 times), and one spectrum may be obtained by averaging the acquired spectral data multiple times.

Although not shown in the flowchart of FIG. 6, if the spectrum data contains a lot of noise, noise may be removed by preprocessing using a filter such as a median or moving average. For the filter processing, it is desirable to select an appropriate filter that matches the characteristics of the noise superimposed on the data.

Next, the spectral intensity normalization unit 12 performs a process of normalizing the intensity of each acquired spectral data (step S12). For example, the intensity of each spectral data is normalized so that the area of the Brillouin scattering spectrum becomes a predetermined value (for example, "1"). This process is performed to eliminate factors such as the material, temperature characteristics, and transmission distance of the optical fiber FUT, and to facilitate comparison of the Brillouin scattering spectra.

Next, a process is performed to find and separate and extract one or more peaks from the normalized spectrum data (step S13: separation and extraction step). If the normalized spectrum data contains a double peak, a process is performed to separate and extract at least two peaks, the maximum peak and the next largest peak.

In this process, for example, a curve represented by a linear sum of two Lorentzian functions is curve-fitted to the sample of spectrum BS as shown in Fig. 3. As a result, a first peak having frequency _f1 and optical intensity _I1 and a second peak having frequency _f2 and optical intensity _I2 are separated and extracted.

The separated and extracted peak information (frequency and intensity) is stored in the peak information storage unit 14 together with the position (distance from one end of the optical fiber FUT) (step S14). The peak information is stored in the form of a database or array, for example, in a magnetic recording medium or semiconductor memory, and can be freely read or rewritten by a program such as software executed on a personal computer (PC).

Next, based on the peak information stored in the peak information storage unit 14, the intensity of the first peak is compared with the intensity of the second peak at each position on the optical fiber FUT, and if they are close to each other, the peak intensity closeness identification unit 15 identifies the position (step S15: identification step). Whether the intensity of the first peak is close to the intensity of the second peak is determined by whether the difference between the intensity of the first peak and the intensity of the second peak is equal to or less than a predefined threshold value.

For example, when the light intensity I is expressed in decibel units such as 5 log ₁₀ (I), if the difference between the intensity of the first peak and the intensity of the second peak is within a threshold value of 0.1 dB, the intensity of the first peak and the intensity of the second peak are determined to be close to each other. This threshold value can be changed depending on, for example, the level of noise in the measuring device of the optical fiber sensor main body 20. The number of spectrum data (the number of positions on the identified optical fiber FUT) in which the intensity of the first peak and the intensity of the second peak are determined to be close to each other is defined as u.

Then, at one of the u locations identified as having close peak intensities, the peak intensity gradient calculation unit 16 calculates the gradient of the peak intensity in the longitudinal direction of the optical fiber FUT (step S16: gradient calculation step). At this time, in addition to the peak intensity at the location where the gradient of the peak intensity is to be calculated, the gradient of the peak intensity is calculated using the peak intensity at the location before and after that location.

The slope of the peak intensity is calculated separately for the low frequency side and the high frequency side. That is, when calculating the slope of the peak intensity on the low frequency side, only the peak intensity on the low frequency side (the intensity of the second peak) is used, and when calculating the slope of the peak intensity on the high frequency side, only the peak intensity on the high frequency side (the intensity of the first peak) is used. By performing this process, for example, the slope S _H of the peak intensity on the high frequency side and the slope S _L of the peak intensity on the low frequency side shown in FIG. 4 are obtained.

Then, the peak determination unit 17 performs a process to determine whether the frequency transition of the maximum peak at the point where the slope of the peak intensity was obtained is a false detection (step S17: determination step). Since the Brillouin scattering spectrum obtained from the correct peak position has a high intensity, it is natural that the transition to the correct peak will be accompanied by a gradual increase in the spectral intensity.

Based on the contents of Table 1 described above, if the frequency transition of the maximum peak is a transition from frequency _fL to frequency _fH and S _H -S _L < 0, it is determined to be a false detection. Also, if the frequency transition of the maximum peak is a transition from frequency _fH to frequency _fL and S _H -S _L > 0, it is determined to be a false detection. If it is determined to be a false detection, the peak information stored in the peak information storage unit 14 is corrected. Specifically, the frequency of the maximum peak when S _H -S _L < 0 is corrected to _fL , and the frequency of the maximum peak when S _H -S _L > 0 is corrected to _fH .

Then, the peak determination unit 17 determines whether the determination in step S17 has been performed for all of the u locations identified in step S15 as having close peak intensities (step S18). If it is determined that the determination has not been performed for all of the u locations (if the determination result is "NO"), the process returns to step S16.

In contrast, if it is determined that the determination has been performed for all u locations (if the determination result is "YES"), the peak determination unit 17 determines that the detection is a false positive, and the peak frequency output unit 18 performs a process of outputting the peak frequency of the maximum peak after the detection is corrected to the correct frequency (step S19). The output here may include not only the Brillouin frequency shift at the u locations identified by the peak intensity proximity identification unit 15 as having close peak intensities, but also the Brillouin frequency shift at all or some of the locations distributed along the length of the optical fiber FUT.

When the above process is completed, the fiber characteristic output unit 19 judges whether or not it is necessary to convert the output peak frequency to temperature or strain (step S20). If it is judged that conversion is necessary (the judgement result is "YES"), the fiber characteristic output unit 19 performs a process of outputting the temperature or strain by multiplying a predetermined conversion coefficient according to the optical fiber FUT (step S21). On the other hand, if it is judged that conversion is not necessary (the judgement result is "NO"), the process of outputting the temperature or strain is not performed, and the series of processes shown in FIG. 6 ends.

The process shown in the flowchart of FIG. 6 may end automatically after processing a predetermined amount of data determined by the user, for example. Alternatively, the process may end after a predetermined number of executions or a predetermined period of time determined by the user. Furthermore, the process may end manually at the user's instruction.

As described above, in this embodiment, first, spectral data showing the spectrum of Brillouin scattered light generated at different positions in the length direction of the optical fiber FUT is obtained, and the first peak and the second peak are separated and extracted for each of the spectral data. Next, the position on the optical fiber where the spectral data in which the intensity of the first peak and the intensity of the second peak are close to each other is identified, and the gradient of the intensity of the first peak and the gradient of the intensity of the second peak at the identified position are calculated. Then, based on the calculated gradient, it is determined which of the first peak and the second peak is appropriate as the maximum peak. This makes it possible to prevent erroneous detection of the maximum peak when two peaks appear in the spectrum of the Brillouin scattered light.

Although the embodiment of the present invention has been described above, the present invention is not limited to the above embodiment and can be freely modified within the scope of the present invention. For example, in the above embodiment, an example was described in which the optical fiber sensor system 1 is composed of a spectrum analyzer 10 and an optical fiber sensor body 20. However, the spectrum analyzer 10 constituting the optical fiber sensor system 1 may be provided within the optical fiber sensor body 20.

In addition, each component provided in the spectrum analysis device 10 can be realized by hardware, software, or a combination of these. Furthermore, the above-mentioned spectrum analysis method is also performed by the cooperation of each component, and can be realized by hardware, software, or a combination of these. Here, being realized by software means being realized by a computer reading and executing a program.

The program can be stored and supplied to a computer using various computer-readable recording media. Computer-readable recording media include, for example, magnetic recording media (e.g., flexible disks, hard disk drives, magneto-optical disks), DVD-ROMs (Read Only Memory), DVD-Rs, DVD-R/Ws, and semiconductor memories (e.g., mask ROMs, PROMs (Programmable ROMs), EPROMs (Erasable PROMs), flash ROMs, and RAMs (random access memories)).

The above-described embodiments may be realized by a single device, or may be realized by cloud computing over a network.

REFERENCE SIGNS LIST 10 Spectral analysis device 11 Spectral data acquisition unit 12 Spectral intensity normalization unit 13 Peak separation and extraction unit 15 Peak intensity proximity identification unit 16 Peak intensity gradient calculation unit 17 Peak determination unit 18 Peak frequency output unit 19 Fiber characteristic output unit FUT Optical fiber

Claims (9)

an acquisition unit that acquires spectral data indicating the spectrum of Brillouin scattered light generated at different positions in the length direction of the optical fiber;
a separation/extraction unit that separates and extracts a first peak and a second peak from each of the spectrum data;
an identifying unit that identifies a position on the optical fiber where the spectrum data in which the intensity of the first peak and the intensity of the second peak are close to each other was obtained;
a gradient calculation unit that calculates a gradient of the intensity of the first peak in the length direction of the optical fiber and a gradient of the intensity of the second peak in the length direction of the optical fiber at the position identified by the identification unit ;
a determination unit that determines whether the first peak or the second peak is a maximum peak based on the gradient calculated by the gradient calculation unit;
A spectrum analysis device comprising: The spectral analysis device according to claim 1, further comprising a normalization unit that normalizes the intensity of each of the spectral data acquired by the acquisition unit. The spectrum analysis device according to claim 1 or 2, wherein the acquisition unit acquires the spectrum data detected by a BOTDR or BOTDA optical fiber sensor. The spectral analysis device according to any one of claims 1 to 3, wherein the separation and extraction unit performs curve fitting using a Lorentz distribution function or a Voigt function having an arbitrary half-width for each of the spectral data to separate and extract the first peak and the second peak. The spectrum analysis device according to any one of claims 1 to 4, wherein the slope calculation unit calculates a slope of the intensity of the first peak by performing a linear regression analysis on the first peak at the position identified by the identification unit and the first peak at positions before and after the position, and calculates a slope of the intensity of the second peak by performing a linear regression analysis on the second peak at the position identified by the identification unit and the second peak at positions before and after the position. The spectrum analysis device according to any one of claims 1 to 5, further comprising a peak frequency output unit that outputs the frequency of a peak determined by the determination unit to be appropriate as a maximum peak. The spectrum analyzer according to claim 6, further comprising a fiber characteristic output unit that outputs the characteristics of the optical fiber determined from the frequency output from the peak frequency output unit. an acquisition step of acquiring spectral data indicating the spectrum of Brillouin scattered light generated at different positions in the length direction of the optical fiber;
a separation and extraction step of separating and extracting a first peak and a second peak from each of the spectrum data;
a step of identifying a position on the optical fiber from which the spectral data, in which the intensity of the first peak and the intensity of the second peak are close to each other, were obtained;
a gradient calculation step of calculating a gradient of the intensity of the first peak in the length direction of the optical fiber and a gradient of the intensity of the second peak in the length direction of the optical fiber at the position specified by the specifying step ;
a determining step of determining whether the first peak or the second peak is appropriate as a maximum peak based on the gradient calculated in the gradient calculating step;
A spectrum analysis method having the following structure: Computer,
an acquisition means for acquiring spectral data indicating the spectrum of Brillouin scattered light generated at different positions in the length direction of the optical fiber;
a separation/extraction means for separating and extracting a first peak and a second peak from each of the spectrum data;
a means for identifying a position on the optical fiber where the spectral data in which the intensity of the first peak and the intensity of the second peak are close to each other was obtained;
a gradient calculation means for calculating a gradient of the intensity of the first peak in the length direction of the optical fiber and a gradient of the intensity of the second peak in the length direction of the optical fiber at the position specified by the specifying means ;
a determining means for determining whether the first peak or the second peak is a maximum peak based on the gradient calculated by the gradient calculating means;
A spectrum analysis program that functions as a

Images