Brillouin signal acquisition system and cable fault positioning and classifying method

Abstract

The invention provides a Brillouin signal acquisition system and a cable fault positioning and classifying method, and belongs to the technical field of power cable intelligent monitoring. The acquisition system comprises an optical path module, a to-be-detected sensing optical fiber module and a backward optical signal receiving and demodulation module which are communicated with one another, the optical path module is used for generating a detection light component and a reference light component, and the detection light component is modulated and then sent to the to-be-detected sensing optical fiber module; the reference light component is sent into a backward light signal receiving and demodulating module; the to-be-detected sensing optical fiber module is used for receiving the detection light component, so that spontaneous Brillouin scattering of the detection light component occurs in the to-be-detected sensing optical fiber, and a backward optical signal is generated and returned to the optical path module; the backward optical signal receiving and demodulating module is used for receiving the reference optical component and the backward optical signal, performing coherent beat frequency on the received reference optical component and backward optical signal, and then collecting to obtain a Brillouin frequency shift signal of the sensing optical fiber module to be detected; according to the method, the fault position and type of the to-be-detected sensing optical fiber module are obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of power cable intelligent monitoring, and in particular to a Brillouin signal acquisition system and a cable fault locating and classifying method. Background Art

[0002] Brillouin Optical Time Domain Reflectometry (BOTDR), with its advantages such as single-ended, long-distance transmission and the ability to measure temperature and strain simultaneously or separately, holds promise as a core technology for cable condition monitoring. Its measurement accuracy directly depends on the signal-to-noise ratio of the Brillouin Gain Spectrum (BGS) and the accuracy of Brillouin Frequency Shift (BFS) extraction. However, the intensity of spontaneous Brillouin scattered light is approximately 50 dB lower than that of the incident light and is susceptible to interference from fiber noise, making it difficult to effectively extract weak fault signals. While existing hardware optimization solutions can partially suppress noise, they cannot address dynamic interference under complex operating conditions.

[0003] The current mainstream method uses digital filtering (Lorentz function fitting) to improve the quality of the Brillouin gain spectrum signal. Low-pass filtering combined with pump wave / probe wave modulation can reduce high-frequency noise, but parameters (such as cutoff frequency and modulation degree) need to be adjusted for different fiber types, resulting in poor algorithm versatility. Although the spectrum reconstruction method based on analytical functions can improve the BFS extraction accuracy to ±1MHz (corresponding to a temperature error of approximately ±1°C and a strain error of approximately ±20°C), the BFS extraction accuracy can be improved to ±1MHz. However, in practice, due to factors such as digitization errors and nonlinear photon-phonon interactions, the deviation in frequency shift extraction can reach over 5 MHz during long-distance monitoring, severely limiting the accuracy of temperature / strain inversion. Traditional machine learning relies on manually constructed statistical models, making it difficult to capture the spatiotemporal propagation characteristics of cable faults. While Spiking Neural Networks (SNNs) have the potential to process spatiotemporal data, their fixed-rate encoding mechanism is unsuitable for the dynamic signal characteristics of BOTDRs, and they lack physical constraints such as cable entities, which affects fault classification accuracy.

[0004] Therefore, it is very necessary to provide a Brillouin signal acquisition system and a cable fault location and classification method to reduce the noise impact of each device on the extracted Brillouin signal, solve the problems of high construction cost and slow response speed of current high-precision Brillouin optical time domain reflectometer instruments, and realize efficient data processing. Summary of the Invention

[0005] In view of this, the present invention proposes a Brillouin signal acquisition system and a cable fault location and classification method based on a pulse neural dynamic spatiotemporal optimization network, which uses software design based on dedicated Brillouin optical time-domain reflectometer hardware conditions to reduce the noise impact of each component on the extracted Brillouin signal, thereby overcoming the shortcomings of current high-precision Brillouin optical time-domain reflectometer instruments, such as difficulty in construction, slow response speed, and low data processing efficiency.

[0006] In one aspect, the present invention provides a Brillouin signal acquisition system, comprising: An optical path module, a sensing optical fiber module to be tested, and a backward optical signal receiving and demodulating module optically connected to each other; The optical path module includes a seed laser and a first fiber coupler. The first fiber coupler divides the continuous light output by the seed laser into a detection light component and a reference light component. The detection light component is modulated and sent to the sensing fiber module to be measured; the reference light component is sent to the backward optical signal receiving and demodulation module. The sensing optical fiber module to be tested is used to receive the detection light component, so that the detection light component undergoes spontaneous Brillouin scattering in the sensing optical fiber module to generate a backward light signal, wherein the backward light signal contains temperature / strain information of the sensing optical fiber module to be tested and returns to the optical path module; The backward optical signal receiving and demodulating module is used to receive the reference light component and the backward optical signal, and collect the Brillouin frequency shift signal of the sensing optical fiber module to be tested after performing coherent beat between the received reference light component and the backward optical signal.

[0007] Based on the above technical solution, preferably, the optical path module further includes a detection light unit, a local oscillator unit and a reference light path unit; the first output end of the first optical fiber coupler is optically connected to the input end of the detection light unit, the second output end of the first optical fiber coupler is optically connected to the input end of the local oscillator unit, the third output end of the first optical fiber coupler is optically connected to the input end of the reference light path unit, and the output end of the local oscillator unit is also optically connected to the input end of the detection light unit; the first optical fiber coupler couples the detection light component to the detection light unit and the local oscillator unit respectively, and when the local oscillator unit receives the detection light component, it also outputs the obtained known Brillouin frequency shift to the detection light unit; the first optical fiber coupler outputs the reference light component to the reference light path unit, and the reference light path unit receives the reference light component and outputs it to the backward optical signal receiving and demodulating module; the intensities of the detection light component or the reference light component outputted from the first output end, the second end and the third end of the first optical fiber coupler are 10%, 10% and 80% of the intensity of the continuous light, respectively.

[0008] Preferably, the local oscillator unit includes a second erbium-doped fiber amplifier, a first optical circulator, and a local fiber Brillouin generator; the second output end of the first optical fiber coupler is optically connected to the input end of the first erbium-doped fiber amplifier, the output end of the first erbium-doped fiber amplifier is optically connected to the first port of the first optical circulator, the second port of the first optical circulator is optically connected to the local fiber Brillouin generator, and the third port of the first optical circulator is optically connected to the detection light unit; the first port and the second port of the first optical circulator are unidirectionally connected, and the second port and the third port of the first optical circulator are unidirectionally connected; the local oscillator unit is used to generate a reference Brillouin signal; The detection light unit includes a second fiber coupler, a second erbium-doped fiber amplifier, and a pulse modulation AOM. The first output end of the first fiber coupler is optically connected to the first input end of the second fiber coupler, the second input end of the second fiber coupler is optically connected to the third port of the first optical circulator, the output end of the second fiber coupler is optically connected to the input end of the second erbium-doped fiber amplifier, and the output end of the second erbium-doped fiber amplifier is optically connected to the input end of the pulse modulation AOM; the output end of the pulse modulation AOM is optically connected to the sensing fiber module to be tested; The reference path optical unit includes a third optical fiber coupler, the third output end of the first optical fiber coupler is optically connected to the first input end of the third optical fiber coupler, the second input end of the third optical fiber coupler is optically connected to the sensing optical fiber module to be measured, and the output end of the third optical fiber coupler is optically connected to the backward optical signal receiving and demodulating module.

[0009] Preferably, the sensing fiber module to be tested includes a second optical circulator and an optical fiber to be tested; the first port of the second optical circulator is connected to the output optical path of the pulse modulation AOM, the second port of the second optical circulator is connected to the optical path of the optical fiber to be tested, and the third port of the second optical circulator is connected to the second input optical path of the third optical fiber coupler; the optical fiber to be tested generates an initial Brillouin frequency shift signal when not affected by the environment, and the initial Brillouin frequency shift signal changes due to the temperature / strain of the optical fiber to be tested, causing the optical fiber to be tested to generate a Brillouin signal; the first port and the second port of the second optical circulator are unidirectionally connected, and the second port and the third port of the second optical circulator are unidirectionally connected.

[0010] Preferably, the backward optical signal receiving and demodulating module includes an integrated coherent receiver and a data collector; the output end of the third optical fiber coupler is optically connected to the input end of the integrated coherent receiver, and the output end of the integrated coherent receiver is electrically connected to the data collector. The integrated coherent receiver is used to obtain the Brillouin frequency shift generated when the optical fiber to be tested is affected by temperature / strain by coherent beat frequency of the received reference optical signal component, the reference Brillouin signal and the Brillouin signal of the optical fiber to be tested. The data collector collects the obtained Brillouin frequency shift and uploads it to the cloud platform.

[0011] On the other hand, the present invention also provides a cable fault location and classification method, comprising the following steps: S1: Configuring the Brillouin signal acquisition system; the optical fiber module to be tested is configured with an optical fiber to be tested, and the optical fiber to be tested is coaxially extended with the cable; S2: The backward optical signal receiving and demodulating module obtains the Brillouin frequency shift signal of the sensor fiber module to be tested and sends it to the host computer. The host computer uploads it to the cloud platform. The cloud platform sequentially executes the Brillouin data preprocessing sub-step, spatiotemporal pulse encoding sub-step, dynamic feature space optimization sub-step, and fault location and classification sub-step, and outputs the results to the cloud platform display interface.

[0012] Preferably, the Brillouin data preprocessing sub-step adopts a hybrid noise reduction method based on multi-scale analysis, which combines the time-frequency localization characteristics of wavelet transform with the dynamic threshold adjustment mechanism to achieve the separation of effective information and noise in the Brillouin frequency shift signal, and then performs data normalization and feature extraction to obtain a feature vector.

[0013] Preferably, the spatiotemporal pulse coding sub-step specifically includes: constructing a pulse neural network, including an input layer, a hidden layer and an output layer, wherein the input layer of the pulse neural network receives an input sequence with unique time characteristics, so that the neurons x State variables Exceeding the threshold θ When a certain time pulse is emitted, the neuron x A series of pulses emitted form a pulse train , , Represents neurons x Issued i The time of a pulse: The connection between the hidden layer and the input layer and the output layer includes multiple synapses. Each synapse corresponds to a different time stage under the same spatial coding unit in the spatiotemporal pulse coding. Each synapse has a different delay and adjustable connection weight. Finally, after processing by several hidden layers, the pulse sequence output by each neuron constitutes the output of the network.

[0014] Preferably, the content of the dynamic feature space optimization sub-step is to use dynamic neural algorithms and cable physical constraints to construct a multi-objective optimization model, define the spatiotemporal distance measurement function of similar fault samples, use the improved Hausdorff distance to calculate the physical position similarity in the spatial dimension, align the pulse sequence through the dynamic time warping algorithm in the time dimension, and introduce the cable topology constraint matrix into the neural dynamic optimization model framework; in view of the pulse timing sensitivity of the pulse neural network, a dual-channel loss function is designed: the first channel minimizes the difference in pulse emission patterns of similar faults, and the second channel maximizes the divergence of different fault types by jointly optimizing the temperature and strain feature weights to achieve the optimal projection of the feature space in the time-frequency domain.

[0015] Preferably, the fault location and classification sub-step is to construct a pulse neural network, configure a positioning model and a classification model, and the positioning model uses a Gaussian mixture model to estimate the probability density of the optimized feature space; the two-dimensional parameters of the fault location are calculated iteratively; the classification model designs a three-channel pulse neural network classifier: a short-time high-frequency pulse cluster corresponds to a short-circuit fault, a medium-frequency continuous pulse cluster represents mechanical damage, and a low-frequency intermittent pulse indicates insulation aging; when more than one classifier unit triggers an abnormal pulse at the same time, the result is returned and a composite judgment of the fault type is performed.

[0016] The Brillouin signal acquisition system and cable fault location and classification method provided by the present invention have the following beneficial effects compared with the prior art: (1) The present invention improves the shortcomings of the existing Brillouin optical time-domain reflectometer, such as large instrument size and low integration, by integrating and optimizing the optical path module and the backward optical signal receiving and demodulating module of the existing Brillouin optical time-domain reflectometer. The present invention is suitable for long-term operation in a semi-enclosed environment, thereby enhancing the reliability of the equipment. (2) The pulse neural dynamic optimization network proposed in the present invention can quickly convert the spatial-temporal information of temperature and strain into temporal pulse frequency and phase when facing the optical cable monitoring data under different working conditions. By adjusting the delay time and gain weight within the synapse, the selective allocation of pulse neurons is achieved, avoiding the influence of the pulse neural dynamic optimization network on the recognition of the cable operation status when too few or too many neurons are set during training; (3) The fault diagnosis module proposed in this invention introduces the physical topology constraints of the cable to achieve multi-objective parameter optimization of the cable, define the time and space distance of similar fault samples, provide a basis for the precise positioning and fault identification of the cable, and effectively suppress the influence of irrelevant parameters on the diagnosis module. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 A schematic diagram of the system structure of a Brillouin signal acquisition system and a cable fault location and classification method according to the present invention; Figure 2 A network flow diagram of a Brillouin signal acquisition system and a cable fault location and classification method according to the present invention; Figure 3This is a schematic diagram of the pulse neural network operation of a Brillouin signal acquisition system and a cable fault location and classification method of the present invention.

[0019] Figure numerals: 1, optical path module; 2, sensing fiber module to be tested; 3, backward optical signal receiving and demodulating module; 0, seed laser; 10, first optical fiber coupler; 11, detection light unit; 12, local oscillator unit; 13, reference path optical unit; 111, second optical fiber coupler; 112, second erbium-doped fiber amplifier; 113, pulse modulation AOM; 121, second erbium-doped fiber amplifier; 122, first optical circulator; 123, local fiber Brillouin generator; 21, second optical circulator; 22, optical fiber to be tested; 31, integrated coherent receiver; 32, data collector. DETAILED DESCRIPTION

[0020] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] The existing Brillouin frequency shift device BOTDR cannot solve dynamic interference; the spectrum reconstruction method based on analytical function is affected by factors such as digitization error and nonlinearity of photon-phonon interaction. The frequency shift deviation is large during long-distance monitoring, which seriously restricts the accuracy of temperature / strain inversion; although the traditional pulse neural network has the potential for spatiotemporal data processing, its fixed rate encoding mechanism cannot adapt to the dynamic signal characteristics of BOTDR, and lacks physical constraint modeling such as cable entities, which affects the accuracy of fault classification. In view of this, Figure 1 As shown, in one aspect, the present invention provides a Brillouin signal acquisition system, comprising: An optical path module 1, a sensing optical fiber module 2 to be tested, and a backward optical signal receiving and demodulating module 3 that are optically connected to each other; The optical path module 1 includes a seed laser 0 and a first optical fiber coupler 10. The first optical fiber coupler 10 divides the continuous light output by the seed laser 0 into a detection light component and a reference light component. The detection light component is modulated and sent to the sensing optical fiber module 2 to be measured; the reference light component is sent to the backward optical signal receiving and demodulating module 3. The sensing optical fiber module 2 to be tested is used to receive the detection light component, so that the detection light component undergoes spontaneous Brillouin scattering in the sensing optical fiber module 2 to generate a backward light signal. The backward light signal contains the temperature / strain information of the sensing optical fiber module 2 to be tested and returns to the optical path module 1. The backward optical signal receiving and demodulating module 3 is used to receive the reference optical component and the backward optical signal, and perform coherent beat between the received reference optical component and the backward optical signal to obtain the Brillouin frequency shift signal of the sensing optical fiber module 2 to be tested.

[0022] like Figure 1 As shown, this embodiment integrates and optimizes the optical path module and the backward optical signal receiving and demodulating module of the existing Brillouin optical time domain reflectometer, thereby improving the problem of the existing Brillouin optical time domain reflectometer being large in size and not easy to carry.

[0023] like Figure 1 As shown, the optical path module 1 also includes a detection light unit 11, a local oscillator unit 12 and a reference light unit 13; the first output end of the first optical fiber coupler 10 is optically connected to the input end of the detection light unit 11, the second output end of the first optical fiber coupler 10 is optically connected to the input end of the local oscillator unit 12, the third output end of the first optical fiber coupler 10 is optically connected to the input end of the reference light unit 13, and the output end of the local oscillator unit 12 is also optically connected to the input end of the detection light unit 11; the first optical fiber coupler 10 couples the detection light component to In the detection light unit 11 and the local oscillator unit 12, when the local oscillator unit 12 receives the detection light component, it also outputs the obtained known Brillouin frequency shift to the detection light unit 11; the first optical fiber coupler 10 outputs the reference light component to the reference path optical unit 13, and the reference path optical unit 13 receives the reference light component and outputs it to the backward optical signal receiving and demodulating module 3; the intensities of the detection light component or the reference light component outputted from the first output end, the second end, and the third end of the first optical fiber coupler 10, respectively, are 10%, 10%, and 80% of the intensity of the continuous light, respectively.

[0024] Specifically, the local oscillator unit 12 includes a second erbium-doped fiber amplifier 121, a first optical circulator 122 and a local fiber Brillouin generator 123. The second output end of the first fiber coupler 10 is optically connected to the input end of the first erbium-doped fiber amplifier 121, the output end of the first erbium-doped fiber amplifier 121 is optically connected to the first port of the first optical circulator 122, the second port of the first optical circulator 122 is optically connected to the local fiber Brillouin generator 123, and the third port of the first optical circulator 122 is optically connected to the detection light unit 11; the first port of the first optical circulator 122 is unidirectionally connected to the second port, and the second port of the first optical circulator 122 is unidirectionally connected to the third port; the local oscillator unit 12 is used to generate a reference Brillouin signal; the local oscillator optical unit 12 is a unit that generates a Brillouin signal by integrating it into the system, and can input a reference Brillouin signal to the backward optical signal receiving and demodulating module 3, and perform coherent beat frequency with the backward Brillouin signal in the optical fiber to be tested.

[0025] The detection light unit 11 includes a second fiber coupler 111, a second erbium-doped fiber amplifier 112, and a pulse modulation AOM 113. The first output end of the first fiber coupler 10 is optically connected to the first input end of the second fiber coupler 111, the second input end of the second fiber coupler 111 is optically connected to the third port of the first optical circulator 122, the output end of the second fiber coupler 111 is optically connected to the input end of the second erbium-doped fiber amplifier 112, and the output end of the second erbium-doped fiber amplifier 112 is optically connected to the input end of the pulse modulation AOM 113; the output end of the pulse modulation AOM 113 is optically connected to the sensing fiber module 2 to be tested; the detection light signal couples the reference Brillouin signal with the detection light component and then sends it to the sensing fiber module 2 to be tested.

[0026] Reference optical path unit 13 includes a third optical fiber coupler. The third output end of first optical fiber coupler 10 is optically connected to the first input end of the third optical fiber coupler. The second input end of the third optical fiber coupler is optically connected to the sensing optical fiber module 2 to be tested. The output end of the third optical fiber coupler is optically connected to the backward optical signal receiving and demodulating module 3. Reference optical path unit 13 is used to provide the backward optical signal and the reference light component to the backward optical signal receiving and demodulating module 3.

[0027] The sensing fiber module 2 to be tested includes a second optical circulator 21 and an optical fiber to be tested 22; the first port of the second optical circulator is optically connected to the output end of the pulse modulation AOM 113, the second port of the second optical circulator 21 is optically connected to the optical fiber to be tested 22, and the third port of the second optical circulator 21 is optically connected to the second input end of the third optical fiber coupler. When the optical fiber to be tested 22 is not affected by the environment, it generates an initial Brillouin frequency shift signal. The initial Brillouin frequency shift signal changes due to the temperature / strain of the optical fiber to be tested 22, causing the optical fiber to be tested 22 to generate a Brillouin signal. The first port of the second optical circulator 21 is unidirectionally connected to the second port, and the second port of the second optical circulator 21 is unidirectionally connected to the third port. It should be noted that the initial Brillouin frequency shift signal generated by the optical fiber to be tested 22 when it is not affected by the environment and the change of the initial Brillouin frequency shift signal due to the temperature / strain of the optical fiber to be tested 22, causing the optical fiber to be tested 22 to generate a Brillouin signal, are both backward optical signals.

[0028] The backward optical signal receiving and demodulating module 3 includes an integrated coherent receiver 31 and a data collector 32; the output end of the third optical fiber coupler is optically connected to the input end of the integrated coherent receiver 31, and the output end of the integrated coherent receiver 31 is electrically connected to the data collector 32. The integrated coherent receiver 31 is used to obtain the Brillouin frequency shift generated when the optical fiber 22 to be tested is affected by temperature / strain by coherent beat frequency of the received reference optical signal component, the reference Brillouin signal and the Brillouin signal of the optical fiber to be tested 22. The data collector 32 collects the obtained Brillouin frequency shift and uploads it to the cloud platform.

[0029] The specific steps for calculating the temperature and strain information of the frequency shift location are as follows: After completing target detection on the spectrum graph, the frequency and distance information are extracted through the coordinate position of the center point of each area, and the relationship between the Brillouin frequency change and temperature and strain is calculated: , where 、 T 0 and are the Brillouin frequency shift, temperature and strain of the optical fiber under initial conditions, and are the Brillouin temperature coefficient and gauge coefficient of the optical fiber, respectively.

[0030] On the other hand, the present invention also provides a cable fault location and classification method, comprising the following steps: S1: Configuring the Brillouin signal acquisition system; the optical fiber module to be tested is configured with an optical fiber to be tested, and the optical fiber to be tested is coaxially extended with the cable; S2: The backward optical signal receiving and demodulating module 3 obtains the Brillouin frequency shift signal of the sensing optical fiber module 2 to be tested and sends it to the host computer, which then uploads it to the cloud platform. Figure 2 As shown, the cloud platform sequentially executes the Brillouin data preprocessing sub-step, spatiotemporal pulse encoding sub-step, dynamic feature space optimization sub-step, and fault location and classification sub-step, and outputs the results to the cloud platform display interface.

[0031] Among them, the Brillouin data preprocessing sub-step adopts a hybrid noise reduction method based on multi-scale analysis. By combining the time-frequency localization characteristics of wavelet transform with the dynamic threshold adjustment mechanism, the effective information and noise in the Brillouin frequency shift signal are separated, and then data normalization and feature extraction are performed; the Brillouin frequency shift signal is decomposed by discrete wavelet, and a wavelet basis function with compact support and regularity is selected. As an analysis tool, it can effectively match the transient characteristics of the Brie, and construct an adaptive threshold function based on the signal energy distribution through the calculated decomposition coefficient : ,in m is the number of wavelet decomposition layers, is the retained low-frequency coefficient, is the noise tolerance threshold; In the data normalization process, a segmented adaptive mapping algorithm based on the physical properties of optical fiber is designed to calculate the nonlinear relationship between Brillouin frequency shift and temperature / strain; the cubic spline interpolation method is used to smooth the curve of the Brillouin frequency shift signal, and then a dual-channel normalization model is established: the temperature parameter T Through the sigmoid function Compressed to the interval [0, 1], the slope parameter αDynamic adjustment according to the fiber type, the strain parameter adopts the arc tangent function Mapped to the range [-1,1], ε For strain, β is the temperature compensation factor; this processing method not only ensures the numerical stability of the data in the pulse encoding stage, but also significantly reduces the storage requirements for subsequent processing.

[0032] In the feature extraction stage, a multimodal anomaly detection architecture is used to design a gradient mutation detection algorithm based on a sliding window in the spatial dimension: the width is set to L 0 initial detection window, Δ l To detect the sliding amount of the window, calculate the temperature gradient in the window = Δ T / Δ l and strain gradient = Δ ε / Δ l The weighted average value is calculated, and the weight coefficient is determined by the Brillouin temperature / gauge coefficient ratio of the optical fiber. When a temperature gradient anomaly is detected and the duration exceeds the set time, the phase continuity of the frequency shift curve is first verified to eliminate false anomalies, and finally the strain gradient mutation criterion is combined to confirm the fault characteristics. Finally, the temperature gradient integral area is extracted from the sliding window. And the energy ratio characteristic Ea / Et of the abnormal area form a characteristic vector; Ea and Et are both areas enclosed by the frequency-shifted signal in a certain space, but the latter is the initial state, and the former is the frequency shift after a certain period of environmental disturbance.

[0033] The spatiotemporal pulse coding sub-step specifically includes: constructing Figure 3 The pulse neural network shown in the figure includes an input layer, a hidden layer and an output layer. The input received by the input layer of the pulse neural network is a pulse sequence with unique time characteristics, which makes the neurons x State variables Exceeding the threshold θ When a certain time pulse is emitted, the neuron x A series of pulses emitted form a pulse train , , Represents neurons x Issued i The connection between the hidden layer and the input and output layers includes multiple synapses. Each synapse corresponds to a different time stage under the same spatial coding unit in the spatiotemporal pulse coding. Each synapse has a different delay and adjustable connection weight. Finally, after processing by several hidden layers, the pulse sequence output by each neuron constitutes the output of the network. It's a neuron j The state variable represents the neuron j Changes in resting potential and action potential under the influence of time series, ,in Represents neurons j The impact on itself after the pulse is generated, Presynaptic neurons i , Represents neurons j The response to the impulse before the synapse is affected, the weight Represents neurons i and neurons j Synapses k The connection strength, represents the delay between different synapses.

[0034] In the spatiotemporal pulse coding stage, an innovative space-time dual coding strategy is designed. In the spatial dimension, each meter of cable is divided into independent coding units, and a neuron model is used to convert the normalized temperature value into a pulse emission frequency: , where f i The perceived temperature T i The corresponding pulse frequency, f 0 is the preset pulse frequency, T max and T min are the maximum and minimum temperature values, respectively. Strain fluctuations are mapped to pulse phase offsets using phase modulation technology. Delayed pulse encoding is implemented for abnormal events detected in the time dimension. Sudden failures correspond to instantaneous high-frequency pulse clusters, while mechanical / chemical damage manifests as sustained low-frequency pulse clusters.

[0035] The content of the dynamic feature space optimization sub-step is to use the dynamic neural algorithm and cable physical constraints to build a multi-objective optimization model, define the time-space distance measurement function of similar fault samples, and use the improved Hausdorff distance in the spatial dimension. Calculating physical location similarity , There are two different points x i and y j The second-order paradigm of the distance is adopted, the pulse sequence is aligned in the time dimension through the dynamic time warping algorithm, and the cable topology constraint matrix is introduced into the neural dynamic optimization model framework; according to the pulse timing sensitivity of the pulse neural network, a dual-channel loss function is designed: the first channel minimizes the difference in pulse emission patterns of similar faults, and the second channel maximizes the divergence of different fault types. By jointly optimizing the temperature and strain feature weights, the optimal projection of the feature space in the time-frequency domain is achieved.

[0036] The fault location and classification sub-steps are to construct a pulse neural network, configure the positioning model and classification model, and use the Gaussian mixture model to estimate the probability density of the optimized feature space in the positioning model; the two-dimensional parameters of the fault location are calculated iteratively; the classification model designs a three-channel pulse neural network classifier: short-term high-frequency pulse clusters correspond to short-circuit faults, medium-frequency continuous pulse clusters represent mechanical damage, and low-frequency intermittent pulses indicate insulation aging; when more than one classifier unit triggers abnormal pulses at the same time, the results are returned and a composite judgment of the fault type is performed.

[0037] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A Brillouin signal acquisition system, characterized in that: include: An optical path module (1), a sensing optical fiber module to be tested (2), and a backward optical signal receiving and demodulating module (3) that are optically connected to each other; The optical path module (1) includes a seed laser (0) and a first optical fiber coupler (10). The first optical fiber coupler (10) divides the continuous light output by the seed laser (0) into a detection light component and a reference light component. The detection light component is modulated and sent to the sensing optical fiber module (2) to be measured; and the reference light component is sent to the backward optical signal receiving and demodulating module (3). The sensing optical fiber module to be tested (2) is used to receive a detection light component, causing the detection light component to undergo spontaneous Brillouin scattering in the sensing optical fiber module to be tested (2), thereby generating a backward light signal, wherein the backward light signal contains temperature / strain information of the sensing optical fiber module to be tested (2) and is returned to the optical path module (1); The backward optical signal receiving and demodulating module (3) is used for receiving the reference optical component and the backward optical signal, and performing coherent beat frequency on the received reference optical component and the backward optical signal to collect the Brillouin frequency shift signal of the sensing optical fiber module (2) to be tested.

2. A Brillouin signal acquisition system according to claim 1, characterized in that: The optical path module (1) further comprises a detection light unit (11), a local oscillator unit (12) and a reference light path unit (13); the first output end of the first optical fiber coupler (10) is optically connected to the input end of the detection light unit (11), the second output end of the first optical fiber coupler (10) is optically connected to the input end of the local oscillator unit (12), the third output end of the first optical fiber coupler (10) is optically connected to the input end of the reference light path unit (13), and the output end of the local oscillator unit (12) is also optically connected to the input end of the detection light unit (11); the first optical fiber coupler (10) splits the detection light into The first optical fiber coupler (10) couples the reference light component to the detection light unit (11) and the local oscillator unit (12); when the local oscillator unit (12) receives the detection light component, it also outputs the obtained known Brillouin frequency shift to the detection light unit (11); the first optical fiber coupler (10) outputs the reference light component to the reference path optical unit (13); the reference path optical unit (13) receives the reference light component and outputs it to the backward optical signal receiving and demodulating module (3); the intensities of the detection light component or the reference light component outputted from the first output end, the second end and the third end of the first optical fiber coupler (10) are 10%, 10% and 80% of the intensity of the continuous light, respectively.

3. A Brillouin signal acquisition system according to claim 2, characterized in that: The local oscillator unit (12) includes a second erbium-doped fiber amplifier (121), a first optical circulator (122), and a local fiber Brillouin generator (123); the second output end of the first optical fiber coupler (10) is optically connected to the input end of the first erbium-doped fiber amplifier (121); the output end of the first erbium-doped fiber amplifier (121) is optically connected to the first port of the first optical circulator (122); the second port of the first optical circulator (122) is optically connected to the local fiber Brillouin generator (123); and the third port of the first optical circulator (122) is optically connected to the detection light unit (11); the first port of the first optical circulator (122) is unidirectionally connected to the second port, and the second port of the first optical circulator (122) is unidirectionally connected to the third port; the local oscillator unit (12) is used to generate a reference Brillouin signal; The detection light unit (11) includes a second optical fiber coupler (111), a second erbium-doped fiber amplifier (112), and a pulse modulation AOM (113); the first output end of the first optical fiber coupler (10) is optically connected to the first input end of the second optical fiber coupler (111); the second input end of the second optical fiber coupler (111) is optically connected to the third port of the first optical circulator (122); the output end of the second optical fiber coupler (111) is optically connected to the input end of the second erbium-doped fiber amplifier (112); and the output end of the second erbium-doped fiber amplifier (112) is optically connected to the input end of the pulse modulation AOM (113); The output end of the pulse modulation AOM (113) is optically connected to the sensing optical fiber module (2) to be tested; The reference optical path unit (13) includes a third optical fiber coupler, a third output end of the first optical fiber coupler (10) is optically connected to a first input end of the third optical fiber coupler, a second input end of the third optical fiber coupler is optically connected to a sensing optical fiber module (2) to be measured, and an output end of the third optical fiber coupler is optically connected to a backward optical signal receiving and demodulating module (3).

4. The Brillouin signal acquisition system according to claim 3, characterized in that: The sensing optical fiber module (2) to be tested comprises a second optical circulator (21) and an optical fiber to be tested (22); a first port of the second optical circulator is optically connected to the output end of the pulse modulation AOM (113), a second port of the second optical circulator (21) is optically connected to the optical fiber to be tested (22), and a third port of the second optical circulator (21) is optically connected to the second input end of the third optical fiber coupler; the optical fiber to be tested (22) generates an initial Brillouin frequency shift signal when not affected by the environment; the initial Brillouin frequency shift signal changes due to the temperature / strain of the optical fiber to be tested (22), so that the optical fiber to be tested (22) generates a Brillouin signal; the first port of the second optical circulator (21) is unidirectionally connected to the second port, and the second port of the second optical circulator (21) is unidirectionally connected to the third port.

5. The Brillouin signal acquisition system according to claim 3, characterized in that: The backward optical signal receiving and demodulating module (3) includes an integrated coherent receiver (31) and a data collector (32); the output end of the third optical fiber coupler is optically connected to the input end of the integrated coherent receiver (31), and the output end of the integrated coherent receiver (31) is electrically connected to the data collector (32). The integrated coherent receiver (31) is used to obtain the Brillouin frequency shift generated when the optical fiber to be tested (22) is affected by temperature / strain by coherent beat frequency using the received reference optical signal component, the reference Brillouin signal and the Brillouin signal of the optical fiber to be tested (22). The data collector (32) collects the obtained Brillouin frequency shift and uploads it to the cloud platform.

6. A cable fault location and classification method, characterized in that: The steps include: S1: configuring the Brillouin signal acquisition system according to any one of claims 1 to 5; the optical fiber to be tested is configured in the optical fiber to be tested, and the optical fiber to be tested is coaxially extended with the cable; S2: The backward optical signal receiving and demodulating module (3) obtains the Brillouin frequency shift signal of the sensing optical fiber module (2) to be tested and sends it to the host computer, which uploads it to the cloud platform. The cloud platform executes the Brillouin data preprocessing sub-step, the spatiotemporal pulse encoding sub-step, the dynamic feature space optimization sub-step, and the fault location and classification sub-step in sequence, and outputs the results to the cloud platform display interface.

7. A cable fault location and classification method according to claim 6, characterized in that: The Brillouin data preprocessing sub-step adopts a hybrid noise reduction method based on multi-scale analysis. By combining the time-frequency localization characteristics of wavelet transform with a dynamic threshold adjustment mechanism, it realizes the separation of effective information and noise in the Brillouin frequency shift signal, and then performs data normalization and feature extraction to obtain a feature vector.

8. A cable fault location and classification method according to claim 7, characterized in that: The spatiotemporal pulse coding sub-step specifically includes: constructing a pulse neural network, including an input layer, a hidden layer, and an output layer. The input layer of the pulse neural network receives an input sequence with unique time characteristics, which makes the neurons x State variables Exceeding the threshold θ When a certain time pulse is emitted, the neuron x A series of pulses emitted form a pulse train , , Represents neurons x Issued i The time of a pulse: The connection between the hidden layer and the input layer and the output layer includes multiple synapses. Each synapse corresponds to a different time stage under the same spatial coding unit in the spatiotemporal pulse coding. Each synapse has a different delay and adjustable connection weight. Finally, after processing by several hidden layers, the pulse sequence output by each neuron constitutes the output of the network.

9. A cable fault location and classification method according to claim 8, characterized in that: The content of the dynamic feature space optimization sub-step is to use the dynamic neural algorithm and cable physical constraints to build a multi-objective optimization model, define the spatiotemporal distance measurement function of similar fault samples, use the improved Hausdorff distance to calculate the physical position similarity in the spatial dimension, align the pulse sequence through the dynamic time warping algorithm in the time dimension, and introduce the cable topology constraint matrix into the neural dynamic optimization model framework; in view of the pulse timing sensitivity of the pulse neural network, a dual-channel loss function is designed: the first channel minimizes the difference in pulse emission patterns of similar faults, and the second channel maximizes the divergence of different fault types by jointly optimizing the temperature and strain feature weights. Achieve the optimal projection of feature space in the time-frequency domain.

10. A cable fault location and classification method according to claim 9, characterized in that: The fault location and classification sub-step involves constructing a pulse neural network, configuring a location model and a classification model. The location model uses a Gaussian mixture model to estimate the probability density of the optimized feature space. The two-dimensional parameters of the fault location are calculated iteratively. The classification model designs a three-channel pulse neural network classifier: short-duration high-frequency pulse clusters correspond to short-circuit faults, medium-frequency continuous pulse clusters represent mechanical damage, and low-frequency intermittent pulses indicate insulation aging. When more than one classifier unit triggers an abnormal pulse at the same time, the result is returned and a composite judgment of the fault type is performed.

Images