Non-intrusive monitoring power optical fiber abnormity identification and positioning method and non-intrusive monitoring power optical fiber abnormity identification and positioning system

Abstract

The invention relates to a non-intrusive monitoring electric power optical fiber abnormity identification and positioning method and system, and the method comprises the following steps: enabling an optical fiber to generate leakage light through a clamping coupler, converting a leakage light signal into an electric signal through a photoelectric detector, converting the photoelectric signal into a digital signal through an analog-to-digital converter, and transmitting the digital signal to a processor; when the digital signal has fluctuation greater than a preset value, marking the optical fiber line as an abnormal line; the optical fiber measurement module is communicated with the abnormal optical fiber line through the optical switch switching module, and a backward Rayleigh scattering signal in the abnormal line is measured; carrying out noise reduction on a back Rayleigh scattering signal in the abnormal line by using wavelet transform; performing abnormal point positioning according to a mutation point on the backward Rayleigh scattering signal after noise reduction, and identifying an abnormal type through a support vector machine model optimized by a grey wolf algorithm; and merging the abnormal points and the abnormal types in the optical fiber line, transmitting the merged abnormal points and abnormal types to a cloud through a data transmission port, and sending out an optical fiber line abnormal early warning.

Description

Technical Field

[0001] The present application relates to the field of optical communication technologies, and more specifically, to a method and system for identifying and locating power fiber anomalies through non-intrusive monitoring. Background Art

[0002] Compared with traditional communication media, optical fibers have the advantages of large capacity, long transmission distance, strong corrosion resistance, and immunity to electromagnetic field interference, so they are widely used in communication transmission systems in power systems. However, with the continuous laying of fiber optic networks, the maintenance and management problems of fiber optic networks have become increasingly prominent. Therefore, the status monitoring of power fibers is very important. The status monitoring of power fibers is crucial for improving the safety, reliability, economy, and intelligence level of power systems. By real-time monitoring and analyzing the operating status of optical fibers, potential problems can be detected in a timely manner and corresponding measures can be taken to ensure the stable operation of power communication systems, guarantee the safety and reliability of power facilities, reduce maintenance costs, and improve emergency response capabilities.

[0003] Traditional methods for monitoring the status of power fibers have many problems such as long repair duration, low automation level, and poor real-time performance, which are not conducive to the rapid elimination of anomalies. Therefore, a new type of power fiber status monitoring method has broad application prospects. In order to reduce the interference of fiber status monitoring on the normal operation of optical fibers, a non-intrusive method can be used to monitor power fibers.

[0004] The prior art, such as the Chinese patent application with the publication number "CN117560074A", discloses an OTDR event recognition method based on signal background noise extraction, including: obtaining a sampling curve of the backscattered signal generated after a preset pulse signal is transmitted through a fiber under test, and determining the sampling resolution and optical blind zone; calculating the sampling fluctuation signal of each point in the sampling curve based on the optical blind zone; performing low-pass filtering on the sampling fluctuation signal to obtain a background noise signal; performing adaptive threshold denoising and hard threshold denoising on the sampling signal according to the background noise signal to obtain an event fluctuation signal; and screening events for the event fluctuation signal according to the threshold corresponding to the preset pulse width to determine whether there are anomalies in the fiber under test.

[0005] The problems existing in the above prior art are that it relies on the OTDR to actively transmit pulse signals, which may interfere with the normal communication of optical fibers; it has poor adaptability to complex noises (such as periodic interference) with low-pass filtering and hard threshold denoising; and it can only judge the type of anomaly (such as splice loss) through the "pulse width threshold" and cannot distinguish specific faults such as flanges and bends. Summary of the Invention

[0006] To solve the above technical problems, the present invention proposes a method and system for identifying and locating power fiber anomalies through non-intrusive monitoring.

[0007] The technical solution of the present invention is as follows:

[0008] The present invention provides a non-invasive monitoring power fiber optic anomaly identification and location system, including:

[0009] A non-invasive optical power monitoring module, an optical switch switching module, an optical fiber measurement module, an optical fiber anomaly identification module, and an anomaly recording module, where:

[0010] The non-invasive optical power monitoring module is used to monitor the optical power value fluctuations in the optical fiber line in real time, and identify the abnormal optical fiber line according to the optical power fluctuations;

[0011] The optical switch switching module is used to connect the optical fiber measurement module to the abnormal optical fiber line;

[0012] The optical fiber measurement module is used to collect the backward Rayleigh scattering signal in the abnormal optical fiber line;

[0013] The optical fiber anomaly identification module is used to locate and classify the abnormal points according to the collected backward Rayleigh scattering signal;

[0014] The anomaly recording module is used to merge the abnormal point location and classification information, upload it to the cloud and issue a warning.

[0015] As a preferred embodiment, the optical fiber measurement module includes: a laser, a circulator, a photodetector, and a acquisition card.

[0016] As a preferred embodiment, the optical fiber anomaly identification module is used to locate and classify the abnormal points according to the collected backward Rayleigh scattering signal. The specific steps are as follows:

[0017] Wavelet transform noise reduction: Use wavelet transform to perform noise reduction processing on the collected backward Rayleigh scattering signal;

[0018] Abnormal location: Calculate the standard deviation of the backward Rayleigh scattering signal data after noise reduction processing through a sliding window. The size of the standard deviation represents the degree of signal mutation. Calculate the standard deviation of all windows, determine the abnormal window of the standard deviation through a preset threshold, and expand to a 100-sampling point window near the abnormal window. Locate the position with the maximum standard deviation as the abnormal occurrence position;

[0019] Signal interception and normalization: Intercept a 101-sampling point window centered on the abnormal occurrence position and perform normalization processing. The normalization formula is:

[0020]

[0021] In the formula: I i is the amplitude of the i-th sampling point intercepted; I i-onr is I i the value after normalization; Imax , I min respectively represent the maximum and minimum amplitudes in the intercepted data;

[0022] Anomaly classification: The 101-dimensional normalized window data is used as the input of the support vector machine model optimized by the trained grey wolf algorithm to obtain the anomaly type.

[0023] As a preferred implementation, for the support vector machine model optimized by the grey wolf algorithm, the penalty parameter C and the kernel function parameter γ in the support vector machine model are optimized by the grey wolf algorithm. The specific steps are as follows:

[0024] Regarding the penalty parameter C and the kernel function parameter γ in the support vector machine model as variables to be optimized, a parameter space is constructed:

[0025] C ∈ [C min , C max ;

[0026] γ ∈ [γ min , γ max ;

[0027] In the formula: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min , C max are respectively the lower and upper limits of the search space of C; γ min , γ max are respectively the lower and upper limits of the search space of γ.

[0028] The fitness function is to maximize the classification accuracy rate, specifically:

[0029] Fitness(X) = 1 - Accuracy cross-valodation (X);

[0030] X = (C, γ);

[0031] In the formula: Fitness(X) is the fitness function value of the grey wolf individual, and the smaller the value, the better the parameter combination X; X is the position vector of the grey wolf individual, a two-dimensional parameter composed of C and γ; Accuracy cross-valodation (X) is the classification accuracy rate of the support vector machine model on the cross-validation set when using the parameter X;

[0032] Randomly generate N grey wolf individuals, and the position X of each grey wolf individual i is evenly distributed within the parameter space;

[0033] For the position X of each grey wolf individual i , calculate the classification accuracy rate of the support vector machine model through k-fold cross-validation to obtain the fitness value:

[0034]

[0035] Where: Fitness i is the fitness of the i-th gray wolf individual; k is the number of folds of cross-validation; TextSet j is the j-th fold validation dataset; Accuracy(X i , TextSet j ) is the classification accuracy of the support vector machine model on the j-th fold test set when using the parameter X i ;

[0036] Select the top three individuals with the best fitness values as the leaders α, β, and δ, and the remaining individuals as ω; each individual ω updates its own position according to the positions of α, β, and δ. The specific steps are as follows:

[0037] Encirclement stage: Calculate the distance vector D from the leader:

[0038] D μ = |C μ ·X μ (t) - X(t)|, where μ ∈ {α, β, δ};

[0039] Where: D μ is the distance vector between the current gray wolf individual and the leader μ; C μ is a random coefficient vector, C μ = 2·rand(0, 1); X μ (t) is the position vector of the leader μ at the t-th iteration;

[0040] Hunting stage: Generate a position offset towards the leader:

[0041]

[0042] Where:

[0043] A μ = 2a·rand(0, 1) - a;

[0044] Where: is the new position vector generated according to the guidance of the leader μ; A μ is a dynamic coefficient vector; a is a convergence factor;

[0045] Position fusion: Integrate the leader's guidance to update the position:

[0046]

[0047] Where: X(t + 1) is the updated position vector of the gray wolf individual at the (t + 1)-th iteration; and New positions calculated based on three leaders α, β, and δ respectively;

[0048] When the maximum number of iterations is reached or the fitness value is lower than the threshold, the optimization is terminated, and the globally optimal parameter combination is output to construct a support vector machine model optimized by the grey wolf algorithm.

[0049] As a preferred embodiment, the convergence factor is adaptively adjusted as the number of iterations increases to balance global exploration and local exploitation. The adaptive adjustment formula of the convergence factor is:

[0050]

[0051] In the formula: a(t) is the value of the convergence factor at the t-th iteration; T max is the maximum number of iterations.

[0052] On the other hand, the present invention also provides a method for non-intrusive monitoring of power optical fiber anomaly identification and location, including the following steps:

[0053] Step S1, make the optical fiber generate leakage light through a clamping coupler, convert the leakage light signal into an electrical signal through a photodetector, and convert the photoelectric signal into a digital signal through an analog-to-digital converter. When the digital signal has a fluctuation greater than a preset value, mark the optical fiber line as an abnormal line;

[0054] Step S2, connect the optical fiber measurement module to the abnormal optical fiber line through an optical switch switching module, and the optical fiber measurement module measures the backward Rayleigh scattering signal in the abnormal line;

[0055] Step S3, use wavelet transform to denoise the backward Rayleigh scattering signal in the abnormal line; locate the abnormal point according to the mutation point on the denoised backward Rayleigh scattering signal, and identify the abnormal type through a support vector machine model optimized by the grey wolf algorithm;

[0056] Step S4, combine the abnormal points and abnormal types in the optical fiber line, transmit them to the cloud through a data transmission port, and issue an optical fiber line anomaly warning.

[0057] As a preferred embodiment, the optical fiber measurement module includes: a laser, a circulator, a photodetector, and an acquisition card.

[0058] As a preferred embodiment, for the support vector machine model optimized by the grey wolf algorithm, the penalty parameter C and the kernel function parameter γ in the support vector machine model are optimized by the grey wolf algorithm. The specific steps are as follows:

[0059] Take the penalty parameter C and the kernel function parameter γ in the support vector machine model as variables to be optimized, and construct a parameter space:

[0060] C∈[Cmin , C max ;

[0061] γ ∈ [γ min , γ max ;

[0062] Where: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min , C max are respectively the lower and upper limits of the search space of C; γ min , γ max are respectively the lower and upper limits of the search space of γ.

[0063] The objective function is to maximize the classification accuracy, that is, to minimize the classification error, specifically:

[0064] Fitness(X) = 1 - Accuracy cross-valodation (X);

[0065] X = (C, γ);

[0066] Where: Fitness(X) is the fitness function value of the gray wolf individual, and the smaller the value, the better the parameter combination X; X is the position vector of the gray wolf individual, a two-dimensional parameter composed of C and γ; Accuracy cross-valodation (X) is the classification accuracy of the support vector machine model on the cross-validation set when using the parameter X;

[0067] Randomly generate N gray wolf individuals, and each gray wolf individual position X i is evenly distributed within the parameter space;

[0068] For each gray wolf individual position X i , calculate the classification accuracy of the support vector machine model through k-fold cross-validation to obtain the fitness value:

[0069]

[0070] Where: Fitness i is the fitness of the i-th gray wolf individual; k is the number of folds of cross-validation; TextSet j is the j-th fold validation data set; Accuracy(X i , TextSet j ) is the classification accuracy of the support vector machine model on the j-th fold test set when using the parameter X i ;

[0071] Select the top three individuals with the best fitness as the leaders α, β, and δ, and the remaining individuals as ω; each individual ω updates its own position according to the positions of α, β, and δ. The specific steps are as follows:

[0072] Encirclement stage: Calculate the distance vector D to the leader:

[0073] D μ = |C μ ·X μ (t) - X(t)|, μ ∈ {α, β, δ};

[0074] In the formula: D μ is the distance vector between the current grey wolf individual and the leader μ; C μ is a random coefficient vector, C μ = 2·rand(0, 1); X μ (t) is the position vector of the leader μ at the t-th iteration;

[0075] Hunting stage: Generate a position offset towards the leader:

[0076]

[0077] Among them:

[0078] A μ = 2a·rand(0, 1) - a;

[0079] In the formula: is the new position vector generated according to the guidance of the leader μ; A μ is a dynamic coefficient vector; a is a convergence factor;

[0080] Position fusion: Update the position comprehensively according to the leader's guidance:

[0081]

[0082] In the formula: X(t + 1) is the updated position vector of the grey wolf individual at the (t + 1)-th iteration; and are the new positions calculated based on the three leaders α, β, and δ respectively.

[0083] When the maximum number of iterations is reached or the fitness value is lower than the threshold, the optimization is terminated, and the global optimal parameter combination is output to construct a support vector machine model optimized by the grey wolf algorithm.

[0084] On the other hand, the present invention also provides an electronic device with a computer program stored thereon, and when the computer program is executed by a processor, it implements a non-intrusive monitoring-based power optical fiber anomaly identification and positioning method as described in any embodiment of the present invention.

[0085] In another aspect, the present invention also provides a computer-readable medium for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement a method for non-invasive monitoring of power fiber anomaly identification and location as described in any embodiment of the present invention.

[0086] The present invention has the following beneficial effects:

[0087] 1. Non-invasive monitoring: The clamping coupler is used to detect the leaked light, avoiding damage to the fiber structure and not affecting the normal communication of the fiber, which is suitable for power communication systems with high reliability requirements;

[0088] 2. High-precision anomaly location: Wavelet transform denoising combined with sliding window standard deviation analysis improves the location accuracy under low signal-to-noise ratio conditions;

[0089] 3. Intelligent classification and optimization: The grey wolf algorithm is used to optimize the SVM parameters, improving the classification accuracy rate compared with traditional methods;

[0090] 4. Real-time warning and cloud integration: Anomaly information is uploaded to the cloud in real time, supporting remote monitoring and rapid maintenance, and reducing the operation and maintenance costs. Description of the Drawings

[0091] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained based on these drawings without creative efforts.

[0092] Figure 1 It is a schematic flow chart of the method of the present invention;

[0093] Figure 2 It is a schematic diagram of the non-invasive detection unit and the optoelectronic conversion unit;

[0094] Figure 3 It is the backward Rayleigh scattering signal measured by the optical fiber measurement module before and after wavelet transform denoising;

[0095] Figure 4 It is the standard deviation of the Rayleigh scattering signal after denoising;

[0096] Figure 5 It is the normalized Rayleigh scattering signal at the anomaly occurrence location;

[0097] Figure 6 It is the layout of the optical fiber to be measured;

[0098] Figure 7 It is a schematic diagram of the non-invasive optical power monitoring module. Detailed implementation manners

[0099] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part rather than all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without making creative efforts shall fall within the protection scope of the present invention.

[0100] It should be understood that the step numbers used in the text are only for convenient description and do not limit the execution order of the steps.

[0101] It should be understood that the terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0102] The terms "comprising" and "including" indicate the presence of the described features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or their combinations.

[0103] The term " / and / " refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0104] Embodiment 1:

[0105] To make the purpose, technical solutions and advantages of the present invention clearer, the following will, with reference to the specific embodiments of the present application and Figure 1 the accompanying drawings, clearly and completely describe the technical solutions of the present invention.

[0106] To solve the problems of the prior art, the present invention provides a method for identifying and locating abnormal conditions of a power optical fiber for non-invasive monitoring, including the following steps:

[0107] Step S1, causing leakage light to be generated in the optical fiber through a clamping coupler, converting the leakage light signal into an electrical signal through a photodetector, and converting the photoelectric signal into a digital signal through an analog-to-digital converter. When the digital signal has a fluctuation greater than a preset value, the optical fiber line is marked as an abnormal line;

[0108] The acquisition of the optical signal is mainly completed by the clamping coupler, which is a special passive device. It causes the optical fiber to have a certain radius of curvature through the clamping groove, resulting in a small amount of light leakage, and then couples the leaked light through the internal optical path and sends it to the photoelectric conversion unit through the output port.

[0109] The optoelectronic conversion unit is mainly responsible for converting the optical signal obtained by bending sensing into an electrical signal, which is achieved through an optoelectronic detector. The schematic diagrams of the non-invasive detection unit and the optoelectronic conversion unit are as Figure 2 shown. Only need to use a clamping coupler to fix a certain section of the power optical fiber, and use an optoelectronic detector to measure the leaked light caused by the bending of the optical fiber. Usually, the bending of the optical fiber will not cause more than 1% loss of optical power, which has almost no impact on the format and transmission quality of the original signal. There is no need to intervene in the power optical cable communication line, and the impact on the optical fiber communication system is very small, even negligible.

[0110] The analog-to-digital conversion unit is to amplify the signal after the optoelectronic detector converts the optical signal into an electrical signal, and then send it into the analog-to-digital converter to convert the optoelectronic signal into a digital signal. When the signal fluctuates by 10%, record the corresponding abnormal line number.

[0111] Step S2, connect the optical fiber measurement module to the abnormal optical fiber line through the optical switch switching module, and the optical fiber measurement module measures the backward Rayleigh scattering signal in the abnormal line;

[0112] The system settings of the optical fiber measurement module are as follows: Use lasers with wavelengths of 1310nm and 1550nm to emit pulsed light respectively, the pulse width is set to 10ns, the measurement time is set to 1min, and all data within the measurement time are subjected to superposition averaging processing.

[0113] Step S3, use wavelet transform to denoise the backward Rayleigh scattering signal in the abnormal line; locate the abnormal points according to the mutation points on the denoised backward Rayleigh scattering signal, and identify the abnormal type through the support vector machine model optimized by the gray wolf algorithm; the specific process is as follows:

[0114] Wavelet transform denoising: Use wavelet transform to denoise the collected backward Rayleigh scattering signal; the function of wavelet transform is selected as the "Symlet" wavelet function, the decomposition layer of wavelet transform is 5 layers, and the vanishing moment of the wavelet function is 8; the Rayleigh scattering signals before and after denoising are as Figure 3 shown.

[0115] Abnormal positioning: Calculate the standard deviation of the data of the backward Rayleigh scattering signal after denoising through a sliding window. The size of the standard deviation represents the mutation degree of the signal. Calculate the standard deviation of all windows, determine the abnormal window of the standard deviation through a preset threshold, and expand it to a 100-sampling-point window near the abnormal window (covering 50 points before and after the abnormality, about 100 meters), and locate the position with the maximum standard deviation as the abnormal occurrence position;

[0116] In this embodiment, 11 consecutive data in the Rayleigh scattering signal are defined as a window (corresponding to about 11 meters of optical fiber length), and use Figure 4The 6,000 data in it can form 5,990 windows (step size is 1), and calculate the standard deviation of all windows in the Rayleigh scattering signal after noise reduction. Whether an anomaly occurs can be judged according to the threshold of the standard deviation. By statistically analyzing the data, the threshold is set to 0.3 a.u. If the standard deviation is greater than this value, the anomaly occurrence position is determined according to the maximum value of 100 surrounding standard deviations to achieve positioning. The wavelet transform makes the curve smoother, and this operation greatly reduces the possibility of misjudgment. Considering that the noise signal is small, when the mean value of 300 consecutive backward Rayleigh scattering signals is less than the threshold, it is suspected to be noise. In this embodiment, the threshold is set to 10 a.u.

[0117] Signal Interception and Normalization: A 101-sampling-point window intercepted with the anomaly occurrence position as the center is normalized. The normalization formula is:

[0118]

[0119] where: I i is the amplitude of the i-th sampling point intercepted; I i-nor is the value after I i is normalized; I max and I min are respectively the maximum and minimum values of the amplitudes in the intercepted data;

[0120] The collected anomaly types include: fiber optic starting point, flange, fiber optic bend, and fiber optic end point, and 101 backward Rayleigh scattering data intercepted with the mutation point as the center. In order to increase the convergence speed, the Rayleigh scattering signal at the anomaly occurrence is normalized. Different anomalies after normalization are as Figure 5 shown.

[0121] Anomaly Classification: The 101-dimensional window data after normalization is used as the input of the support vector machine model optimized by the trained grey wolf algorithm to obtain the anomaly type.

[0122] In order to reduce the probability of system misjudgment, the present invention uses the grey wolf algorithm to optimize the support vector machine for training. For this purpose, an experiment is designed to obtain a large number of backward Rayleigh scattering signals. The optical fiber measurement module is respectively connected to the optical fibers to be measured in various arrangements, and continuously collects the Rayleigh scattering signals in the optical fibers. The arrangements of the optical fibers to be measured are as Figure 6 shown, and the data diversity is increased by replacing flange devices multiple times, adjusting the tightness of the flange, and adjusting the degree of bending.

[0123] The support vector machine model optimized by the grey wolf algorithm optimizes the penalty parameter C and the kernel function parameter γ in the support vector machine model. The specific steps are as follows:

[0124] Take the penalty parameter C and the kernel function parameter γ in the support vector machine model as variables to be optimized, and construct a parameter space:

[0125] C ∈ [C min , C max = [0.01, 20];

[0126] γ ∈ [γ min , γ max = [0.001, 20];

[0127] Where: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min , C max are respectively the lower and upper limits of the search space of C; γ min , γ max are respectively the lower and upper limits of the search space of γ.

[0128] The fitness function is to maximize the classification accuracy, specifically:

[0129] Fitness(X) = 1 - Accuracy cross-valodation (X);

[0130] X = (C, γ);

[0131] Where: Fitness(X) is the fitness function value of the gray wolf individual, and the smaller the value, the better the parameter combination X; X is the position vector of the gray wolf individual, a two-dimensional parameter composed of C and γ; Accuracy cross-valodation (X) is the classification accuracy of the support vector machine model on the cross-validation set when using the parameter X;

[0132] Randomly generate N gray wolf individuals, and the position X of each gray wolf individual i is evenly distributed within the parameter space;

[0133] For the position X of each gray wolf individual i , calculate the classification accuracy of the support vector machine model through k-fold cross-validation to obtain the fitness value:

[0134]

[0135] Where: Fitness i is the fitness of the i-th gray wolf individual; k is the number of folds of cross-validation; TextSet j is the j-th fold validation data set; Accuracy(X i , TextSet j ) is the classification accuracy of the support vector machine model on the j-th fold test set when using the parameter X i ;

[0136] Select the three optimal individuals according to the fitness value as leaders α, β, and δ, and the remaining individuals are ω; each individual ω updates its own position according to the positions of α, β, and δ. The specific steps are as follows:

[0137] Encirclement phase: Calculate the distance vector D from the leader:

[0138] D μ = |C μ ·X μ (t) - X(t)|, μ ∈ {α, β, δ};

[0139] In the formula: D μ is the distance vector between the current gray wolf individual and the leader μ; C μ is a random coefficient vector, C μ = 2·rand(0, 1); X μ (t) is the position vector of the leader μ at the t-th iteration;

[0140] Hunting phase: Generate a position offset towards the leader:

[0141]

[0142] Among them:

[0143] A μ = 2a·rand(0, 1) - a;

[0144] In the formula: is the new position vector generated according to the guidance of the leader μ; A μ is a dynamic coefficient vector; a is the convergence factor;

[0145] Position fusion: Update the position comprehensively under the guidance of the leader:

[0146]

[0147] In the formula: X(t + 1) is the updated position vector of the gray wolf individual at the (t + 1)-th iteration; and are the new positions calculated based on the three leaders α, β, and δ respectively;

[0148] When the maximum number of iterations is reached or the fitness value is lower than the threshold, terminate the optimization, output the global optimal parameter combination, and construct a support vector machine model optimized by the gray wolf algorithm.

[0149] The convergence factor is adaptively adjusted as the number of iterations increases to balance global exploration and local exploitation. The adaptive adjustment formula of the convergence factor is:

[0150]

[0151] Where: a(t) is the value of the convergence factor at the t-th iteration; T max is the maximum number of iterations, taking 50.

[0152] SVMs are respectively established by using the default penalty parameter and kernel function parameter of SVM and the optimized penalty parameter and kernel function parameter by PSO and GWO; among them, the penalty parameter of SVM is set to 1, and the kernel function parameter is determined by automatic optimization in MATLAB; the penalty parameter and kernel function parameter optimized by PSO are 11.199 and 4.435; the penalty parameter and kernel function parameter optimized by GWO are 16.601 and 4.684 respectively. The abnormal classification accuracies of the three classifiers are shown in Table 1.

[0153] Table 1 Abnormal recognition accuracies of three classifiers

[0154] Classifier Classification accuracy SVM 97.4％ PSO - SVM 98.72％ GWO - SVM 99.36％

[0155] It can be obviously found that the recognition rate of GWO-SVM proposed by the present invention is the highest, reaching 99.36%, while those of SVM and PSO-SVM are 97.4% and 98.72% respectively. It can be seen that GWO-SVM can more effectively realize the abnormal recognition in Rayleigh scattering signals, thereby reflecting the superiority of the present invention.

[0156] Step S4, merge the abnormal points and abnormal types in the optical fiber line, and transmit them to the cloud through the data transmission port and issue an early warning of optical fiber line abnormality.

[0157] Embodiment 2:

[0158] Non-invasive optical power monitoring module, optical switch switching module, optical fiber measurement module, optical fiber abnormality recognition module and abnormality recording module, where:

[0159] The non-invasive optical power monitoring module is used to monitor the fluctuation of the optical power value in the optical fiber line in real time, and identify the abnormal optical fiber line according to the optical power fluctuation;

[0160] The optical switch switching module is used to connect the optical fiber measurement module with the abnormal optical fiber line;

[0161] The optical fiber measurement module is used to collect the backward Rayleigh scattering signal in the abnormal optical fiber line;

[0162] The optical fiber abnormality recognition module is used to locate and classify the abnormal points according to the collected backward Rayleigh scattering signal;

[0163] The abnormality recording module is used to merge the abnormality point location and classification information, upload it to the cloud and issue an early warning.

[0164] Such as Figure 7As shown in the figure, the non-invasive optical power monitoring module includes: a non-invasive optical signal detection unit, a photoelectric conversion unit, and an analog-to-digital conversion unit; among them, the non-invasive optical signal detection unit bends the optical fiber through the clamping groove of the clamping coupler, causing optical leakage, and after coupling through the internal optical path, it is sent to the photoelectric conversion unit through the output port; the photoelectric conversion unit converts the acquired optical signal into an electrical signal through a photodetector and transmits it to the analog-to-digital conversion unit; the analog-to-digital conversion unit converts the electrical signal into a digital signal representing the optical power value. When the digital signal representing the optical power value fluctuates greater than the preset value, the corresponding optical fiber line is marked as an abnormal line.

[0165] The optical fiber measurement module includes: a laser, a circulator, a photodetector, and a data acquisition card.

[0166] The optical fiber anomaly identification module is used to locate and classify the anomaly points according to the collected backward Rayleigh scattering signals. The specific steps are as follows:

[0167] Wavelet transform noise reduction: Use wavelet transform to perform noise reduction processing on the collected backward Rayleigh scattering signals;

[0168] Anomaly localization: Calculate the standard deviation of the data of the backward Rayleigh scattering signals after noise reduction processing through a sliding window. The size of the standard deviation represents the degree of signal mutation. Calculate the standard deviation of all windows, determine the abnormal window of the standard deviation through a preset threshold, and expand it to a 100-sampling-point window near the abnormal window. Locate the position with the maximum standard deviation as the anomaly occurrence position;

[0169] Signal interception and normalization: Intercept a 101-sampling-point window centered on the anomaly occurrence position and perform normalization processing. The normalization formula is:

[0170]

[0171] In the formula: I i is the amplitude of the i-th sampling point intercepted; I i-nor is the value after I i is normalized; I max and I min are respectively the maximum and minimum amplitudes in the intercepted data;

[0172] Anomaly classification: Use the 101-dimensional window data after normalization processing as the input of the support vector machine model optimized by the trained grey wolf algorithm to obtain the anomaly type.

[0173] The support vector machine model optimized by the grey wolf algorithm optimizes the penalty parameter C and the kernel function parameter γ in the support vector machine model. The specific steps are as follows:

[0174] Take the penalty parameter C and the kernel function parameter γ in the support vector machine model as the variables to be optimized, and construct the parameter space:

[0175] C ∈ [C min , C max ;

[0176] γ ∈ [γ min , γ max ;

[0177] Where: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min , C max are respectively the lower and upper limits of the search space of C; γ min , γ max are respectively the lower and upper limits of the search space of γ.

[0178] The fitness function is to maximize the classification accuracy, specifically:

[0179] Fitness(X) = 1 - Accuracy cross-valodation (X);

[0180] X = (C, γ);

[0181] Where: Fitness(X) is the fitness function value of the grey wolf individual, and the smaller the value, the better the parameter combination X; X is the position vector of the grey wolf individual, a two-dimensional parameter composed of C and γ; Accuracy cross-valodation (X) is the classification accuracy of the support vector machine model on the cross-validation set when using the parameter X;

[0182] Randomly generate N grey wolf individuals, and the position X of each grey wolf individual i is evenly distributed within the parameter space;

[0183] For the position X of each grey wolf individual i , calculate the classification accuracy of the support vector machine model through k-fold cross-validation, and obtain the fitness value:

[0184]

[0185] Where: Fitness i is the fitness of the i-th grey wolf individual; k is the number of folds of cross-validation; TextSet j is the j-th fold validation data set; Accuracy(X i , TextSet j ) is the classification accuracy of the support vector machine model on the j-th fold test set when using the parameter X i ;

[0186] Select the top three individuals with the best fitness values as leaders α, β, and δ, and the remaining individuals as ω; each individual ω updates its own position according to the positions of α, β, and δ. The specific steps are as follows:

[0187] Encirclement stage: Calculate the distance vector D from the leader:

[0188] D μ = |C μ ·X μ (t) - X(t)|, μ ∈ {α, β, δ};

[0189] In the formula: D μ is the distance vector between the current grey wolf individual and the leader μ; C μ is a random coefficient vector, C μ = 2·rand(0, 1); X μ (t) is the position vector of the leader μ at the t-th iteration;

[0190] Hunting stage: Generate a position offset towards the leader:

[0191]

[0192] Among them:

[0193] A μ = 2a·rand(0, 1) - a;

[0194] In the formula: is the new position vector generated according to the guidance of the leader μ; A μ is a dynamic coefficient vector; a is the convergence factor;

[0195] Position fusion: Integrate the leader's guidance to update the position:

[0196]

[0197] In the formula: X(t + 1) is the updated position vector of the grey wolf individual at the (t + 1)-th iteration; and are the new positions calculated based on the three leaders α, β, and δ respectively;

[0198] When the maximum number of iterations is reached or the fitness value is lower than the threshold, terminate the optimization, output the global optimal parameter combination, and construct a support vector machine model optimized by the grey wolf algorithm.

[0199] The convergence factor is adaptively adjusted as the number of iterations increases to balance global exploration and local exploitation. The adaptive adjustment formula of the convergence factor is:

[0200]

[0201] Where: a(t) is the value of the convergence factor at the t-th iteration; T max is the maximum number of iterations.

[0202] The upload port of the abnormal record module is the Secure Shell port, and the uploaded information includes: the abnormal occurrence time, the abnormal occurrence location, and the abnormal type.

[0203] Embodiment 3:

[0204] This embodiment provides an electronic device on which a computer program is stored. When the computer program is executed by a processor, it implements a non-intrusive monitoring-based power fiber abnormal identification and positioning method as described in any embodiment of the present invention.

[0205] Embodiment 4:

[0206] This embodiment provides a computer-readable medium for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors are caused to implement a non-intrusive monitoring-based power fiber abnormal identification and positioning method as described in any embodiment of the present invention.

[0207] In the embodiments of the present application, "at least one" means one or more, and "a plurality" means two or more. "And / or" describes the association relationship of associated objects and indicates that three relationships may exist. For example, A and / or B may represent the case where A exists alone, A and B exist simultaneously, or B exists alone. Where A and B may be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after. "At least one of the following" and its similar expressions refer to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, and c may represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c may be single or multiple.

[0208] Those of ordinary skill in the art can realize that the units and algorithm steps described in the embodiments disclosed herein can be implemented by a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the present application.

[0209] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0210] In several embodiments provided by the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or a part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (hereinafter referred to as ROM), random access memory (hereinafter referred to as RAM), magnetic disks, or optical discs that can store program codes.

[0211] The above are only the embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A non-invasive monitoring power optical fiber anomaly identification and positioning system, characterized in that: include: Non-intrusive optical power monitoring module, optical switch module, optical fiber measurement module, optical fiber anomaly identification module and anomaly recording module, wherein: Non-intrusive optical power monitoring module, used to monitor the optical power fluctuation in the optical fiber line in real time, and identify abnormal optical fiber lines according to the optical power fluctuation; An optical switch module, used to connect the optical fiber measurement module with the abnormal optical fiber line; Optical fiber measurement module, used to collect backscattered Rayleigh signals in abnormal optical fiber lines; The optical fiber anomaly identification module is used to locate and classify anomalies based on the collected backscattered Rayleigh scattering signals; The anomaly recording module is used to merge the anomaly location and classification information, upload it to the cloud and issue an early warning.

2. The non-invasive monitoring power optical fiber anomaly identification and positioning system according to claim 1 is characterized in that: The optical fiber measurement module includes: a laser, a circulator, a photoelectric detector and an acquisition card.

3. The non-invasive monitoring power optical fiber anomaly identification and positioning system according to claim 1, characterized in that: The optical fiber anomaly identification module is used to locate and classify abnormal points according to the collected backscattered Rayleigh scattering signals. The specific steps are as follows: Wavelet transform denoising: Use wavelet transform to denoise the collected Rayleigh backscattering signal; Anomaly location: The standard deviation of the Rayleigh backscattering signal data after noise reduction is calculated through a sliding window. The size of the standard deviation represents the degree of mutation of the signal. The standard deviation of all windows is calculated, and the standard deviation abnormal window is determined by a preset threshold. The window is expanded to 100 sampling points near the abnormal window, and the position with the maximum standard deviation is located as the position where the anomaly occurs; Signal capture and normalization: A 101-sampling point window is captured with the abnormality location as the center and normalized. The normalization formula is: Where: I i is the amplitude of the i-th sampling point cut out; I i-nor For I i Normalized value; I max ,I min They are respectively the maximum and minimum amplitude values ​​in the intercepted data; Anomaly classification: The 101-dimensional normalized window data is used as the input of the support vector machine model optimized by the trained gray wolf algorithm to obtain the anomaly type.

4. The non-intrusive monitoring power optical fiber anomaly identification and positioning system according to claim 3 is characterized by: The support vector machine model optimized by the gray wolf algorithm, the penalty parameter C and the kernel function parameter γ in the support vector machine model are optimized by the gray wolf algorithm, and the specific steps are as follows: The penalty parameter C and kernel function parameter γ in the support vector machine model are taken as variables to be optimized, and the parameter space is constructed: C∈[C min ,C max ]; γ∈[γ min ,c max ]; Where: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min ,C max are the lower and upper bounds of the search space of C respectively; γ min ,γ max It is divided into the lower and upper bounds of the search space of γ. The fitness function is to maximize the classification accuracy, specifically: Fitness(X)=1-Accuracy cross-valodation (X); X = (C, γ); Where: Fitness(X) is the fitness function value of the gray wolf individual. The smaller the value, the better the parameter combination X. X is the position vector of the gray wolf individual, a two-dimensional parameter composed of C and γ. Accuracy cross-valodation (X) is the classification accuracy of the support vector machine model on the cross-validation set when using parameter X; Randomly generate N gray wolf individuals, each gray wolf individual position X i Uniformly distributed in parameter space; For each gray wolf individual position X i , calculate the classification accuracy of the support vector machine model through k-fold cross validation and get the fitness value: Where: Fitness i is the fitness of the i-th gray wolf individual; k is the number of cross-validation folds; TextSet j is the j-fold validation dataset; Accuracy(X i ,TextSet j ) is the parameter X i When , the classification accuracy of the support vector machine model on the j-fold test set; The best three individuals are selected as leaders α, β and δ according to the fitness value, and the remaining individuals are ω; each individual ω updates its own position according to the positions of α, β and δ. The specific steps are as follows: Encirclement phase: Calculate the distance vector D to the leader: D μ =|C μ ·X μ (t)-X(t)|,μ∈{α,β,δ}; Where: D μ is the distance vector between the current gray wolf individual and the leader μ; C μ is the random coefficient vector, C μ =2·rand(0,1);X μ (t) is the position vector of the leader μ at the tth iteration; Hunting phase: Spawn position offset towards leader: in: A μ =2a·rand(0,1)-a; Where: A is the new position vector generated according to the guidance of the leader μ; μ is the dynamic coefficient vector; a is the convergence factor; Position Fusion: Integrated Leaders Guide Update Position: Where: X(t+1) is the updated position vector of the gray wolf at the t+1th iteration; and are the new positions calculated based on the three leaders α, β and δ respectively; When the maximum number of iterations is reached or the fitness value is lower than the threshold, the optimization is terminated, the global optimal parameter combination is output, and the support vector machine model optimized by the Grey Wolf Algorithm is constructed.

5. The non-intrusive monitoring power optical fiber anomaly identification and positioning system according to claim 4, characterized in that: The convergence factor is adaptively adjusted as the number of iterations increases to balance global exploration and local development. The adaptive adjustment formula of the convergence factor is: Where: a(t) is the value of the convergence factor when the number of iterations is t; T max is the maximum number of iterations.

6. A non-invasive monitoring method for identifying and locating power optical fiber anomalies, characterized in that: The following steps are involved: Step S1, using a clamping coupler to cause the optical fiber to generate leakage light, converting the leakage light signal into an electrical signal through a photoelectric detector, and converting the photoelectric signal into a digital signal through an analog-to-digital converter, and marking the optical fiber line as an abnormal line when the digital signal fluctuates more than a preset value; Step S2, connecting the optical fiber measurement module to the abnormal optical fiber line through the optical switch switching module, and the optical fiber measurement module measures the backscattered Rayleigh signal in the abnormal line; Step S3, using wavelet transform to reduce noise on the Rayleigh backscattering signal in the abnormal line; locating the abnormal point according to the mutation point on the Rayleigh backscattering signal after noise reduction, and identifying the abnormal type through the support vector machine model optimized by the gray wolf algorithm; Step S4, the abnormal points and abnormal types in the optical fiber line are merged, transmitted to the cloud through the data transmission port, and an optical fiber line abnormality warning is issued.

7. The method for identifying and locating anomalies of power optical fibers by non-invasive monitoring according to claim 6, characterized in that: The optical fiber measurement module includes: a laser, a circulator, a photoelectric detector and an acquisition card.

8. The method for identifying and locating abnormalities of power optical fibers by non-invasive monitoring according to claim 6, characterized in that: The support vector machine model optimized by the gray wolf algorithm, the penalty parameter C and the kernel function parameter γ in the support vector machine model are optimized by the gray wolf algorithm, and the specific steps are as follows: The penalty parameter C and kernel function parameter γ in the support vector machine model are taken as variables to be optimized, and the parameter space is constructed: C∈[C min ,C max ]; γ∈[γ min ,c max ]; Where: C is the penalty parameter of the support vector machine; γ is the kernel function parameter of the support vector machine; C min ,C max are the lower and upper bounds of the search space of C respectively; γ min ,γ max It is divided into the lower and upper bounds of the search space of γ. The objective function is to maximize the classification accuracy, that is, to minimize the classification error, specifically: Fitness(X)=1-Accuracy cross-valodation (X); X = (C, γ); Where: Fitness(X) is the fitness function value of the gray wolf individual. The smaller the value, the better the parameter combination X. X is the position vector of the gray wolf individual, a two-dimensional parameter composed of C and γ. Accuracy cross-valodation (X) is the classification accuracy of the support vector machine model on the cross-validation set when using parameter X; Randomly generate N gray wolf individuals, each gray wolf individual position X i Uniformly distributed in parameter space; For each gray wolf individual position X i , calculate the classification accuracy of the support vector machine model through k-fold cross validation and get the fitness value: Where: Fitness i is the fitness of the i-th gray wolf individual; k is the number of cross-validation folds; TextSet j is the j-fold validation dataset; Accuracy(X i ,TextSet j ) is the parameter X i When , the classification accuracy of the support vector machine model on the j-fold test set; The best three individuals are selected as leaders α, β and δ according to the fitness value, and the remaining individuals are ω; each individual ω updates its own position according to the positions of α, β and δ. The specific steps are as follows: Encirclement phase: Calculate the distance vector D to the leader: D μ =|C μ ·X μ (t)-X(t)|,μ∈{α,β,δ}; Where: D μ is the distance vector between the current gray wolf individual and the leader μ; C μ is the random coefficient vector, C μ =2·rand(0,1);X μ (t) is the position vector of the leader μ at the tth iteration; Hunting phase: Spawn position offset towards leader: in: A μ =2a·rand(0,1)-a; Where: A is the new position vector generated according to the guidance of the leader μ; μ is the dynamic coefficient vector; a is the convergence factor; Position Fusion: Integrated Leaders Guide Update Position: Where: X(t+1) is the updated position vector of the gray wolf individual at the t+1th iteration; and These are the new positions calculated based on the three leaders α, β and δ respectively. When the maximum number of iterations is reached or the fitness value is lower than the threshold, the optimization is terminated, the global optimal parameter combination is output, and the support vector machine model optimized by the Grey Wolf Algorithm is constructed.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for identifying and locating anomalies of power optical fiber using non-invasive monitoring as described in any one of claims 6 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a method for identifying and locating abnormalities of a power optical fiber by non-invasive monitoring as described in any one of claims 6 to 8 is implemented.

Images