Cable defect detection method and device

Abstract

The invention discloses a cable defect detection method and device, and relates to the field of power equipment fault detection, and the method comprises the steps: generating N electric signals with different frequencies in a preset frequency range, inputting the electric signals into a target cable, testing the input impedance of the target cable under the N electric signals with different frequencies, the broadband impedance of the target cable in the frequency range is obtained; according to the broadband impedance, generating an impedance amplitude and an impedance phase of the target cable at each preset frequency point, and constructing an impedance matrix; inputting the impedance matrix into a fault classification model, and identifying and outputting the category of the target cable through the fault classification model; and if the target cable has a fault, extracting the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable, and determining the fault position of the target cable through a fault positioning model according to the resonant frequency with the maximum amplitude and the typical wave velocity of the target cable. According to the invention, the fault of the target cable can be automatically, quickly and accurately identified, and the fault position can be positioned.

Description

Technical Field

[0001] The present application relates to the field of power equipment fault detection, and in particular to a cable defect detection method and device. Background Art

[0002] Power cables often develop local defects due to long-term use or environmental influences. These defects can degrade power cable performance and, in severe cases, cause power system failures. Therefore, accurately and quickly detecting and locating local defects in power cables is crucial.

[0003] Among existing fault detection methods, the direct current method determines short circuit or overload by detecting whether the current exceeds a threshold, but it cannot locate the fault. The DC test method (such as the insulation resistance test) can only determine whether there is leakage in a certain section of the line, and also cannot locate the fault. Summary of the Invention

[0004] The purpose of this application is to provide a cable defect detection method and device to solve the problem that existing fault detection methods cannot locate the fault location.

[0005] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a method for detecting local defects in a cable, comprising: Generating N electrical signals of different frequencies within a preset frequency range and inputting them into a target cable, testing the input impedance of the target cable under the electrical signals of the N different frequencies, and obtaining the broadband impedance of the target cable within the frequency range; According to the broadband impedance of the target cable, the impedance amplitude and impedance phase of the target cable at each preset frequency point are generated, and the impedance matrix X∈R is constructed. N×2 ; Inputting the impedance matrix X into the constructed fault classification model, identifying and outputting the category of the target cable through the fault classification model; wherein the category includes fault or normal; If the target cable is faulty, the resonant frequency with the largest amplitude in the broadband impedance amplitude spectrum of the target cable is extracted, and the fault location of the target cable is determined by constructing a fault location model based on the resonant frequency with the largest amplitude and the typical wave velocity of the target cable.

[0006] Optionally, the fault classification model is constructed according to the following steps: Construct a training dataset: According to the constructed cable distribution parameter equivalent model, the transmission line equation is established; Determine an electrical parameter calculation formula at any frequency based on the transmission line equation, and calculate the electrical parameters per unit length of each sample cable at any frequency using the electrical parameter calculation formula; Correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode; Within a preset frequency range, N frequency points are set for each sample cable according to the preset frequency range and frequency scanning step, and the input impedance of the head end of each sample cable under each fault mode and each frequency point in the normal mode is calculated according to the electrical parameters per unit length of each sample cable under any frequency in the normal mode; Based on the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode, the impedance amplitude and impedance phase of the head end of each sample cable at each frequency point in each fault mode and normal mode are generated, and broadband impedance spectrum data of each sample cable at N frequency points in each fault mode and normal mode are obtained to form a training data set; According to the training data set, a CNN-based classification model is trained to obtain the fault classification model.

[0007] Optionally, the correction rule includes a first correction rule, a second correction rule and a third correction rule, wherein: The first amendment rule includes: R fault = R k R , k R ∈[0.001,0.1]; The second revised rule includes: C fault = C k C , k C ∈[0.2,0.5]; The third revised rule includes: G aging = G k σ , k σ ∈[10,1000]; Among them, R fault is the corrected resistance per unit length, C fault is the corrected capacitance per unit length, G aging is the corrected conductance per unit length, R The resistance per unit length is expressed asG It represents the insulation conductivity per unit length, C Represents capacitance per unit length, k R is the resistance correction coefficient, k C is the capacitance correction factor, k σ is the conductance correction factor.

[0008] Optionally, the failure mode includes short circuit failure, partial discharge failure and insulation aging failure; The step of correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode specifically includes: Correcting the resistance per unit length of each of the sample cables using the first correction rule while maintaining the inductance, conductance, and capacitance per unit length of each of the sample cables unchanged, thereby obtaining electrical parameters per unit length of each of the sample cables at any frequency of a short-circuit fault; Correcting the capacitance per unit length of each of the sample cables using the second correction rule while keeping the inductance, conductance, and resistance per unit length of each of the sample cables unchanged, thereby obtaining electrical parameters per unit length of each of the sample cables at any frequency of a partial discharge fault; The conductance per unit length of each of the sample cables is corrected by the third correction rule, while the inductance, capacitance, and resistance per unit length of each of the sample cables are kept unchanged, thereby obtaining the electrical parameters per unit length of each of the sample cables at any frequency of insulation aging failure.

[0009] Optionally, the step of constructing a training data set further includes: Before correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule, each sample cable is divided into M segments from the head end to the terminal end of each sample cable; Correcting the resistance per unit length of each of the sample cables by using the first correction rule includes: Starting from different segments of the M segments of each sample cable, the same k R And the resistance per unit length of the subsequent segments is corrected by the first correction rule to simulate short circuit faults at different locations; and / or, starting from the same segment of the M segments of each sample cable, different k R and correcting the resistance per unit length of the subsequent segments using the first correction rule to simulate short-circuit faults of different short-circuit degrees; Maintaining the capacitance, inductance, and conductance per unit length of each section of each sample cable and the resistance per unit length of other sections unchanged; The second correction rule is used to correct the capacitance per unit length of each sample cable, including: The capacitance per unit length of each sample cable is measured using different k C And the correction is performed by the second correction rule to simulate partial discharge faults with different air gap sizes; The conductance per unit length of each sample cable is corrected by the third correction rule, including: The conductance per unit length of each sample cable is measured using different k σ The second correction rule is used to perform correction to simulate insulation aging faults with different insulation aging degrees.

[0010] Optionally, calculating the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode according to the electrical parameters per unit length of each sample cable at any frequency in each fault mode and normal mode specifically includes: According to the following formula, starting from the Mth segment of each sample cable at each frequency point in each fault mode or normal mode, the impedance input impedance of the first segment of each sample cable at each frequency point in each fault mode and normal mode is calculated recursively from back to front. : ; ; ; in, The first frequency of each sample cable under any fault mode or normal mode i The input impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i The characteristic impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i +1 segment input impedance, The first frequency of each sample cable under any fault mode or normal mode i The propagation constant of the segment, is the length of each section of the sample cable, 、 、 、 Corresponding to the first frequency point of each sample cable in any fault mode or normal mode i The resistance, inductance, conductance and capacitance of the segment, , is the imaginary unit, represents the angular frequency,ω = 2 πf , f Indicates the frequency of any frequency point.

[0011] Optionally, the fault classification model includes a CNN network and a fault classifier, wherein: The fault classifier is used to identify the category of the target cable according to the input overall feature vector, the category includes fault or normal, and the fault includes short circuit fault, partial discharge fault, and insulation aging fault; The CNN network includes: The first convolutional layer is used to extract local features of the impedance matrix X to obtain local frequency domain features; A maximum pooling layer is used to reduce the dimension of the local frequency domain features to obtain the local frequency domain features after dimensionality reduction; The second convolutional layer is used to extract deep features of the local frequency domain features after dimensionality reduction to obtain deep frequency domain features; The global average pooling layer is used to perform a global average pooling operation on the deep frequency domain features and output the overall feature vector.

[0012] Optionally, the fault classifier includes: A first fault classifier is used to identify the category of the target cable as a short circuit fault or normal according to the overall feature vector; A second fault classifier is used to identify the category of the target cable as a partial discharge fault or a normal condition based on the overall feature vector; The third classification model is used to identify the category of the target cable as insulation aging fault or normal according to the overall feature vector.

[0013] Optionally, the fault location model includes: ; in, is the fault distance, is the typical wave speed preset according to the cable type, is the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable.

[0014] In a second aspect, the present application provides a cable local defect detection device, comprising: An impedance testing module is configured to generate N electrical signals of different frequencies within a preset frequency range and input them into a target cable, thereby testing the input impedance of the target cable under the electrical signals of N different frequencies and obtaining the broadband impedance of the target cable within the frequency range; The impedance matrix construction module is used to generate the impedance amplitude and impedance phase of the target cable at each preset frequency point according to the broadband impedance of the target cable, and construct the impedance matrix X∈R N×2 ; A fault identification module is configured to input the impedance matrix X into the constructed fault classification model, identify and output the category of the target cable through the fault classification model; wherein the category includes fault or normal; The fault location module is used to extract the resonant frequency with the largest amplitude in the broadband impedance amplitude spectrum of the target cable if the target cable is faulty, and determine the fault location of the target cable through the constructed fault location model based on the resonant frequency with the largest amplitude and the typical wave velocity of the target cable.

[0015] According to the specific embodiments provided in this application, this application discloses the following technical effects: The present application provides a method and apparatus for detecting local defects in a cable. The method obtains the broadband impedance of a target cable through testing, generates the impedance amplitude and impedance phase of the target cable at each frequency point based on the broadband impedance of the target cable, and then constructs an impedance matrix to comprehensively reflect the electrical characteristics of different fault types. The impedance matrix X is input into a constructed fault classification model, and the fault classification model is combined with the impedance matrix X to identify and output the fault category of the target cable, thereby automatically identifying the fault of the target cable. If the target cable is faulty, the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable is extracted, and a fault location model is constructed based on the resonant frequency with the maximum amplitude and the typical wave velocity of the target cable. This determines the fault location of the target cable, thereby locating the fault location of the target cable and resolving the problem that existing fault detection methods are unable to locate the fault location. In addition, the fault classification model is used to automatically identify the faults of the target cable. Compared with the existing technology, there is no need to manually rely on expert experience to determine the fault type and fault location of the target cable, which effectively improves the efficiency of local cable defect detection. At the same time, because the impedance matrix X contains multi-dimensional frequency impedance information (impedance amplitude + impedance phase), it can fully reflect the electrical characteristics of different fault types. Therefore, the fault classification model combined with the impedance matrix X can improve the accuracy of identifying the fault type of the target cable. In summary, the present application realizes the automatic, rapid and accurate identification of faults in the target cable and the location of the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 A schematic diagram of a flow chart of a cable local defect detection method provided in one embodiment of the present application; Figure 2 A schematic diagram of a cable distributed parameter equivalent model provided in an embodiment of the application; Figure 3 A schematic diagram of the structure of a fault classification model provided in another embodiment of the present application; Figure 4 A schematic diagram of the functional modules of a cable local defect detection device provided in one embodiment of the present application; Figure 5 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0019] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0020] In an exemplary embodiment, Figure 1 As shown, a method for detecting local defects in a cable is provided, which is executed by a computer device and includes the following steps 101 to 104. In which: Step 101: Generate N electrical signals of different frequencies within a preset frequency range and input them into a target cable, test the input impedance of the target cable under the electrical signals of N different frequencies, and obtain the broadband impedance of the target cable within the frequency range.

[0021] In the embodiments of the present application, the target cable refers to the cable to be tested. A network analyzer is used to perform a frequency sweep test on the target cable within a set frequency range, measuring the complex input impedance of the target cable at each frequency point. The broadband impedance of the target cable over the frequency range includes the input impedance of the target cable under electrical signals of N different frequencies.

[0022] The number of frequency points N is determined by the preset frequency range and step size. If the preset frequency range lower limit is f min , the upper limit of the frequency range is f max , the preset step size is Δf, then f min is the frequency value of the first frequency point, N is (fmax -f min ) / Δf is rounded down, and the frequency of the second frequency point is f min +Δf, and so on, the frequency value of the next frequency point is obtained by moving Δf based on the frequency value of the previous frequency point, until the frequency value of the next frequency point is f min +(N-1)Δf is not less than f max , at this time, with f max As the frequency value of the last frequency point. For example, if f min =0Hz, f max =10MHz, Δf=1kHz, then N=1000. N determines the dimension and fineness of the impedance matrix.

[0023] Step 102: Generate the impedance amplitude and impedance phase of the target cable at each frequency point based on the broadband impedance of the target cable, and construct an impedance matrix X.

[0024] In the embodiment of the present application, X∈R N×2 Each row of the impedance matrix represents the impedance magnitude and impedance phase at the corresponding frequency point. The impedance matrix X contains multi-dimensional frequency impedance information (impedance magnitude + impedance phase), which can fully reflect the electrical characteristics of different fault types.

[0025] Step 103: Input the impedance matrix X into the constructed fault classification model, and identify and output the fault category of the target cable through the fault classification model.

[0026] In the embodiment of the present application, the impedance matrix X is used in combination with the fault classification model to identify different faults, which effectively improves the recognition and generalization capabilities of the fault classification model for different faults.

[0027] Step 104 : If the target cable is faulty, extract the resonance frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable, and determine the fault location of the target cable using the constructed fault location model based on the extracted resonance frequency with the maximum amplitude and the typical wave velocity of the target cable.

[0028] In the present embodiment, the impedance amplitude of the target cable at each frequency point is arranged by frequency to form a broadband impedance amplitude spectrum. The typical wave velocity of the target cable is determined by the type of target cable. Specifically, the wave velocity of the target cable is determined by the dielectric constant of its insulation material. The typical wave velocity value of the target cable can be obtained by looking up a table or empirically.

[0029] By implementing the above steps 101 to 104, the broadband impedance of the target cable is obtained through testing, and the impedance amplitude and impedance phase of the target cable at each frequency point are generated based on the broadband impedance of the target cable, thereby constructing an impedance matrix to comprehensively reflect the electrical characteristics of different fault types. By inputting the impedance matrix X into the constructed fault classification model, the fault classification model is combined with the impedance matrix X to identify and output the fault type of the target cable, thereby automatically identifying the fault of the target cable. If the target cable is faulty, the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable is extracted, and the fault location model is constructed based on the resonant frequency with the maximum amplitude and the typical wave velocity of the target cable, thereby determining the fault location of the target cable. This achieves fault location of the target cable, solving the problem that existing fault detection methods cannot locate the fault location. In addition, the fault classification model is used to automatically identify the faults of the target cable. Compared with the existing technology, there is no need to manually rely on expert experience to determine the fault type and fault location of the target cable, which effectively improves the efficiency of local cable defect detection. At the same time, because the impedance matrix X contains multi-dimensional frequency impedance information (impedance amplitude + impedance phase), it can fully reflect the electrical characteristics of different fault types. Therefore, the fault classification model combined with the impedance matrix X can improve the accuracy of identifying the fault type of the target cable. In summary, the present application realizes the automatic, rapid and accurate identification of faults in the target cable and the location of the fault.

[0030] In another exemplary embodiment of the present application, the above-mentioned fault classification model is constructed according to the following steps 201 to 202. In which: Step 201: construct a training data set.

[0031] In the embodiment of the present application, a training data set is constructed according to the following steps 21 to 225. In which: Step 2021: Establish a transmission line equation based on the constructed cable distribution parameter equivalent model.

[0032] In the embodiment of the present application, the cable distribution parameter equivalent model constructed is as follows Figure 2 As shown, it is an equivalent circuit unit based on transmission line theory. Figure 2 In, Δ l Indicates unit length (micro cable length), I ( x ) indicates that the x The current at V ( x ) indicates that the x Voltage at position x is the starting point of the unit length cable, Δ I Represents the current change per unit length, Δ V It represents the voltage change per unit length.R Indicates the resistance per unit length (Ω / m), L Indicates the inductance per unit length (H / m), G Indicates the insulation conductivity per unit length (S / m), C Indicates capacitance per unit length (F / m).

[0033] according to Figure 2 , the transmission line equations constructed include: ; ; in, represents the imaginary unit, represents the angular frequency, ω = 2 πf , f Indicates frequency.

[0034] Step 2022: Determine an electrical parameter calculation formula at any frequency based on the transmission line equation, and calculate the electrical parameters per unit length of each sample cable at any frequency based on the electrical parameter calculation formula.

[0035] In the embodiments of the present application, the aforementioned electrical parameters include resistance, inductance, capacitance, and conductance. The sample cables are fault-free cables. Therefore, the calculated electrical parameters per unit length of each sample cable at any frequency are the electrical parameters per unit length of the sample cable at any frequency in normal mode.

[0036] Step 2023: According to a preset correction rule, at least one electrical parameter per unit length of each sample cable at any frequency is corrected to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode.

[0037] In step 2024, within a preset frequency range, N frequency points are set for the sample cables according to the preset frequency range and the frequency scanning step Δf. Based on the electrical parameters per unit length of each sample cable at any frequency point in each fault mode and the normal mode, the input impedance of the head end of each sample cable at each frequency point in each fault mode and the normal mode is calculated.

[0038] Step 2025: Generate the impedance amplitude and impedance phase of the head end of each sample cable at each frequency point in each fault mode and normal mode based on the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode, and obtain broadband impedance spectrum data of each sample cable at N frequency points in each fault mode and normal mode to form a training data set.

[0039] In the embodiment of the present application, the size of the broadband impedance spectrum data of each sample cable at N frequency points in any fault mode or normal mode is N×2, N rows correspond to N frequency points, and each row of each broadband impedance spectrum data is the impedance amplitude and impedance phase of the head end of the sample cable at the corresponding frequency point in any fault mode or normal mode.

[0040] There is no specific limit on the amount of data in the training dataset and it can be set according to actual needs. For example, using multiple sample cables, 500 pieces of short-circuit fault data at different locations, 500 pieces of partial discharge fault data with different air gap sizes, 500 pieces of aging fault data, and 500 pieces of normal mode data can be set.

[0041] Step 202: Train a CNN-based classification model based on the constructed training data set to obtain a fault classification model.

[0042] In an embodiment of the present application, a CNN-based classification model includes a CNN network and a classifier. The CNN network is used to extract features from the impedance matrix of an input sample cable in a certain fault mode or normal mode, and the classifier is used to predict the input category based on the features extracted by the CNN network. During the training process, the training data set is divided into a training set and a test set. Each broadband impedance spectrum data in the training set is input into the CNN-based classification model. The CNN-based classification model predicts the category label corresponding to each broadband impedance spectrum data. Based on the predicted category label and the actual category label, the hyperparameters are continuously adjusted to ensure that the prediction performance of the CNN-based classification model meets the requirements. Finally, the trained CNN-based classification model is tested on the test set. If the test results do not meet the requirements, the hyperparameters of the CNN-based classification model are continuously adjusted until the prediction performance of the CNN-based classification model meets the requirements, thereby obtaining a fault classification model.

[0043] In another exemplary embodiment of the present application, according to the above transmission line equation, the electrical parameter calculation formula obtained includes: ; ; ; ; in, is the vacuum permeability, represents the angular frequency, ω = 2 πf , f Indicates frequency, is the core radius of the sample cable, is the inner radius of the shielding layer of the sample cable, is the core resistivity of the sample cable, is the shield resistivity of the sample cable, is the dielectric constant of the shielding layer of the sample cable, is the core conductivity of the sample cable.

[0044] In another exemplary embodiment of the present application, in the above step 2023, the correction rule includes a first correction rule, a second correction rule, and a third correction rule, wherein: The First Amendment rules include: R fault = R k R , k R ∈[0.001,0.1]; The second amendment rules include: C fault = C k C , k C ∈[0.2,0.5]; The third amendment rules include: G aging = G k σ , k σ ∈[10,1000]; Among them, R fault is the corrected resistance per unit length, C fault is the corrected capacitance per unit length, G aging is the corrected conductance per unit length, k R is the resistance correction coefficient, k C is the capacitance correction factor, k σ is the conductance correction factor.

[0045] In another exemplary embodiment of the present application, the failure modes in the above step 2023 include short circuit failure, partial discharge failure and insulation aging failure.

[0046] Accordingly, the above step 2023 includes: By using the first correction rule, the resistance per unit length of each sample cable is corrected, while the inductance, conductance, and capacitance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of the short-circuit fault; By using the second correction rule, the capacitance per unit length of each sample cable is corrected while the inductance, conductance, and resistance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of the partial discharge fault; The conductance per unit length of each sample cable is corrected by the third correction rule, while the inductance, capacitance, and resistance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of insulation aging failure.

[0047] In the embodiment of the present application, based on the design of fault physical mechanism and feature separability, each fault mode only modifies one dominant electrical parameter.

[0048] In another exemplary embodiment of the present application, the above step 201 further includes: Before step 2023, each sample cable is divided into M segments from the head end to the terminal end of each sample cable.

[0049] Short circuit faults include short circuit faults of different locations and / or short circuit degrees, partial discharge faults include partial discharge faults of different air gap sizes, and insulation aging faults include insulation aging faults of different insulation aging degrees.

[0050] At this time, the resistance per unit length of each sample cable is corrected by the first correction rule, while the inductance, conductance, and capacitance per unit length of each sample cable are kept unchanged, including: Starting from different segments of the M segments of each sample cable, the same k R And the resistance per unit length of the subsequent segments is corrected by the first correction rule to simulate short circuit faults at different locations; and / or, starting from the same segment of the M segments of each sample cable, different k R The resistance per unit length of the subsequent segment is corrected by the first correction rule to simulate short-circuit faults of different short-circuit degrees; The capacitance, inductance, and conductance per unit length of each cable segment and the resistance per unit length of the other segments were kept constant for each sample cable.

[0051] The capacitance per unit length of each sample cable is corrected by the second correction rule, while the inductance, conductance, and resistance per unit length of each sample cable are kept unchanged, including: The capacitance per unit length of each sample cable is measured using different k C (k C The larger the gap, the smaller the partial discharge fault) and corrected by the second correction rule to simulate partial discharge faults with different air gap sizes.

[0052] The conductance per unit length of each sample cable is corrected using the third correction rule, including: The conductance per unit length of each sample cable is measured using different k σ (The greater the insulation aging, the greater the k σThe larger the value, the better) and corrected by the second correction rule to simulate insulation aging faults with different insulation aging degrees.

[0053] In another exemplary embodiment of the present application, in the above step 2024, the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode is calculated based on the electrical parameters per unit length of each sample cable at any frequency in each fault mode and normal mode, including: According to the following formula, starting from the Mth segment of each sample cable at each frequency point in each fault mode or normal mode, recursively calculate the input impedance of the first segment of each sample cable at each frequency point in each fault mode and normal mode from back to front: : ; ; ; in, The first frequency of each sample cable under any fault mode or normal mode i The input impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i The characteristic impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i +1 segment input impedance, The first frequency of each sample cable under any fault mode or normal mode i The propagation constant of the segment, is the length of each section of the sample cable, = l / N, l is the total length of the sample cable; 、 、 、 The corresponding value is the first frequency point of each sample cable under any fault mode or normal mode. i The resistance, inductance, conductance and capacitance of the segment, .

[0054] In the embodiment of the present application, the terminal impedance Z of each sample cable is initialized. L =∞ (indicates that the far end of the sample cable is open).

[0055] In another exemplary embodiment of the present application, it can be seen from the above that the fault classification model includes a CNN network and a fault classifier, wherein the fault classifier is used to identify the category of the target cable based on the input overall feature vector, the category includes fault or normal, and the fault includes short circuit fault, partial discharge fault, and insulation aging fault.

[0056] like Figure 3 As shown, the CNN network includes: The first convolutional layer is used to extract local features of the impedance matrix X and obtain local frequency domain features; The maximum pooling layer is used to reduce the dimension of the local frequency domain features and obtain the local frequency domain features after dimensionality reduction; The second convolutional layer is used to extract the deep features of the local frequency domain features after dimensionality reduction to obtain deep frequency domain features; The global average pooling layer is used to perform global average pooling operations on deep frequency domain features and output the overall feature vector.

[0057] In another exemplary embodiment of the present application, the first convolutional layer has 64 filters with a kernel length of 5, a step size of 1, and an activation function of ReLU. The window size of the maximum pooling layer is 2, and the step size is 2. The second convolutional layer uses 128 filters with a kernel length of 3, a step size of 1, and an activation function of ReLU to capture deep frequency domain features (finer granularity, variation features across multiple frequency points, such as local frequency perturbations, weak resonances, etc., which help identify complex fault modes). The global average pooling layer outputs a 128-dimensional feature vector h, using a fully connected approach with a total of 128 neurons, and the activation function uses ReLU for subsequent fault classification.

[0058] In another exemplary embodiment of the present application, the fault classifier includes: The first fault classifier is used to identify the category of the target cable as a short circuit fault or normal according to the overall feature vector output by the global average pooling layer; The second fault classifier is used to identify the category of the target cable as partial discharge fault or normal according to the overall feature vector output by the global average pooling layer; The third classification model is used to identify the category of the target cable as insulation aging fault or normal based on the overall feature vector output by the global average pooling layer.

[0059] In an embodiment of the present application, the overall feature vector output by the global average pooling layer is simultaneously input into the first fault classifier, the second fault classifier, and the third fault classifier, and the first fault classifier, the second fault classifier, and the third fault classifier are used to identify short-circuit faults, partial discharge faults, and insulation aging faults. If there is a short-circuit fault in the cable, the categories output by the second fault classifier and the third classification model are both normal, and only the category output by the first fault classifier is a fault. If there is a partial discharge fault in the cable, the categories output by the first fault classifier and the third classification model are both normal, and only the category output by the second fault classifier is a fault. If there is an insulation aging fault in the cable, the categories output by the first fault classifier and the second classification model are both normal, and only the category output by the third fault classifier is a fault.

[0060] In another exemplary embodiment of the present application, the first fault classifier, the second fault classifier, and the third classification model adopt SVM (Support Vector Machine) classifiers.

[0061] In the embodiment of the present application, the kernel function of the SVM classifier adopts the Gaussian kernel function. Compared with other classifiers, the SVM classifier effectively improves the nonlinear processing capability and calculation speed.

[0062] CNNs can automatically learn effective features from raw impedance spectra, eliminating the need for manual design of feature extraction rules (such as frequency selection and gradient calculation), thus avoiding complex feature engineering. SVMs perform classification based on the high-dimensional features extracted by CNNs, eliminating the need for manually defined discrimination boundaries and thus reducing overall reliance on prior knowledge and experience. Therefore, combining CNNs with SVM classifiers reduces feature engineering and reliance on prior knowledge.

[0063] In another exemplary embodiment of the present application, the above-mentioned fault classification model further outputs the confidence level of the fault category of the target cable.

[0064] In the embodiments of this application, confidence is typically calculated based on the decision distance output by the SVM. The larger the distance, the more confident the fault classification model is that the prediction is correct. Specifically, the closer the distance is to the center of a particular class, the higher the confidence. A sigmoid function is used to normalize the decision distance output by the SVM to a probabilistic confidence level between 0 and 1 for multi-class comparison and reliability assessment.

[0065] In another exemplary embodiment of the present application, in step 104 above, the fault location model includes: ; in, The fault distance is the straight-line distance from the starting end of the target cable (the end where the electrical signal is injected) to the fault point. is the typical wave speed preset according to the cable type, It is the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable.

[0066] In the embodiment of the present application, since the resonance order n=1 is the strongest and clearest fundamental frequency mode, in order to improve positioning reliability and simplify calculation, the fault location is calculated according to the fundamental frequency mode (resonance order n=1).

[0067] Based on the same inventive concept, embodiments of the present application also provide a cable local defect detection device for implementing the aforementioned cable local defect detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more embodiments of the cable local defect detection device provided below can be found in the above-described limitations of the cable local defect detection method and will not be further elaborated here.

[0068] In an exemplary embodiment, Figure 4 As shown, a cable local defect detection device 30 is provided, comprising: The impedance testing module 301 is configured to generate N electrical signals of different frequencies within a preset frequency range and input them into a target cable, thereby testing the input impedance of the target cable under the electrical signals of the N different frequencies and obtaining the broadband impedance of the target cable within the frequency range. The impedance matrix construction module 302 is used to generate the impedance amplitude and impedance phase of the target cable at each preset frequency point according to the broadband impedance of the target cable, and construct the impedance matrix X∈R N×2 ; The fault identification module 303 is used to input the impedance matrix X into the constructed fault classification model, identify and output the category of the target cable through the fault classification model; wherein the category includes fault or normal; The fault location module 304 is configured to extract the resonance frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable if the target cable is faulty, and determine the fault location of the target cable using the constructed fault location model based on the resonance frequency with the maximum amplitude and the typical wave velocity of the target cable.

[0069] In the embodiment of the present application, the target cable refers to the cable to be tested. A frequency sweep test is performed on the target cable within a set frequency range using a network analyzer to measure the complex input impedance of the target cable at each frequency point.

[0070] The number of frequency points N is determined by the preset frequency range and step size. If the preset frequency range lower limit is f min , the upper limit of the frequency range is f max , the preset step size is Δf, then f min is the frequency value of the first frequency point, N is (f max -fmin ) / Δf is rounded down, and the frequency of the second frequency point is f min +Δf, and so on, the frequency value of the next frequency point is obtained by moving Δf based on the frequency value of the previous frequency point, until the frequency value of the next frequency point is f min +(N-1)Δf is not less than f max , at this time, with f max As the frequency value of the last frequency point. For example, if f min =0Hz, f max =10MHz, Δf=1kHz, then N=1000. N determines the dimension and fineness of the impedance matrix.

[0071] Each row of the impedance matrix represents the impedance magnitude and phase at the corresponding frequency point. The impedance matrix X contains multidimensional frequency-dependent impedance information (impedance magnitude and impedance phase), comprehensively reflecting the electrical characteristics of different fault types. Using the impedance matrix X in conjunction with the fault classification model to identify different faults effectively improves the fault classification model's ability to identify and generalize different faults.

[0072] By arranging the impedance amplitude of the target cable at each frequency point by frequency, a broadband impedance amplitude spectrum can be generated. The typical wave velocity of the target cable is determined by the type of cable. Specifically, the wave velocity of the target cable is determined by the dielectric constant of its insulation material. The typical wave velocity value of the target cable can be obtained by looking up the table or empirically.

[0073] The broadband impedance of the target cable is obtained through testing. The impedance amplitude and impedance phase of the target cable at each frequency point are generated based on the broadband impedance of the target cable, and then an impedance matrix is constructed to comprehensively reflect the electrical characteristics of different fault types. The impedance matrix X is input into the constructed fault classification model, and the fault classification model is combined with the impedance matrix X to identify and output the fault category of the target cable, thereby achieving automatic identification of the target cable fault. If the target cable is faulty, the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable is extracted. Based on the maximum amplitude resonant frequency and the typical wave velocity of the target cable, a fault location model is constructed to determine the fault location of the target cable, thereby achieving fault location of the target cable and solving the problem that existing fault detection methods cannot locate the fault location. In addition, the fault classification model is used to automatically identify the faults of the target cable. Compared with the existing technology, there is no need to manually rely on expert experience to determine the fault type and fault location of the target cable, which effectively improves the efficiency of local cable defect detection. At the same time, because the impedance matrix X contains multi-dimensional frequency impedance information (impedance amplitude + impedance phase), it can fully reflect the electrical characteristics of different fault types. Therefore, identifying the fault type of the target cable through the fault classification model combined with the impedance matrix X can improve the accuracy of identifying the fault type of the target cable. In summary, the present application realizes the automatic, rapid and accurate identification of the fault type of the target cable and the location of the fault.

[0074] In another exemplary embodiment of the present application, the fault identification module 303 is further configured to: Construct a training dataset: According to the constructed cable distribution parameter equivalent model, the transmission line equation is established; According to the above transmission line equation, determine the electrical parameter calculation formula at any frequency, and calculate the electrical parameters per unit length of each sample cable at any frequency according to the electrical parameter calculation formula; Correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode; Within a preset frequency range, N frequency points are set for the sample cables according to the preset frequency range and frequency scanning step Δf. Based on the electrical parameters per unit length of each sample cable at any frequency point in each fault mode and normal mode, the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode is calculated; Based on the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode, the impedance amplitude and impedance phase of the head end of each sample cable at each frequency point in each fault mode and normal mode are generated, and broadband impedance spectrum data of each sample cable at N frequency points in each fault mode and normal mode are obtained to form a training data set; According to the constructed training data set, a CNN-based classification model is trained to obtain a fault classification model.

[0075] In the embodiment of the present application, the cable distribution parameter equivalent model constructed is as follows Figure 2 As shown, it is an equivalent circuit unit based on transmission line theory. Figure 2 In, Δ l Indicates unit length (micro cable length), I ( x ) indicates that the x The current at V ( x ) indicates that the x Voltage at position x is the starting point of the unit length cable, Δ I Represents the current change per unit length, Δ V It represents the voltage change per unit length. R Indicates the resistance per unit length (Ω / m), L Indicates the inductance per unit length (H / m), G Indicates the insulation conductivity per unit length (S / m), C Indicates capacitance per unit length (F / m).

[0076] according to Figure 2 , the transmission line equations constructed include: ; ; in, represents the imaginary unit, represents the angular frequency, ω = 2 πf , f Indicates frequency.

[0077] The sample cables are cables without faults. Therefore, the calculated electrical parameters per unit length of each sample cable at any frequency are the electrical parameters per unit length of the sample cable at any frequency in the normal mode.

[0078] The size of the broadband impedance spectrum data of each sample cable in any fault mode or normal mode is N×2, N rows correspond to N frequency points, and each row of each broadband impedance spectrum data is the impedance amplitude and impedance phase of the head end of the sample cable in any fault mode or normal mode at the corresponding frequency point.

[0079] The CNN-based classification model consists of a CNN network and a classifier. The CNN network is used to extract features from the impedance matrix of the input sample cable in a certain fault mode or normal mode, and the classifier is used to predict the input category based on the features extracted by the CNN network. During the training process, the training data set is divided into a training set and a test set. Each broadband impedance spectrum data point in the training set is input into the CNN-based classification model, which predicts the corresponding category label for each broadband impedance spectrum data point. Based on the predicted category label and the actual category label, the hyperparameters of the CNN-based classification model are continuously adjusted to ensure that the prediction performance of the CNN-based classification model meets the requirements. Finally, the trained CNN-based classification model is tested on the test set. If the test results do not meet the requirements, the hyperparameters of the CNN-based classification model are further adjusted until the prediction performance of the CNN-based classification model meets the requirements, thus obtaining the fault classification model.

[0080] The relevant introductions to electrical parameters, electrical parameter calculation formulas, correction rules, failure modes, and failure classification models are detailed in the description of the above method embodiments and will not be repeated here.

[0081] In another exemplary embodiment of the present application, the failure modes include short circuit failure, partial discharge failure and insulation aging failure.

[0082] Accordingly, the above-mentioned fault identification module 303 is further used to: By using the first correction rule, the resistance per unit length of each sample cable is corrected, while the inductance, conductance, and capacitance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of the short-circuit fault; By using the second correction rule, the capacitance per unit length of each sample cable is corrected while the inductance, conductance, and resistance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of the partial discharge fault; The conductance per unit length of each sample cable is corrected by the third correction rule, while the inductance, capacitance, and resistance per unit length of each sample cable are kept unchanged, thereby obtaining the electrical parameters per unit length of each sample cable at any frequency of insulation aging failure.

[0083] In another exemplary embodiment of the present application, the fault identification module 303 is further configured to: Before correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule, each sample cable is divided into M segments from the head end to the terminal end.

[0084] In another exemplary embodiment of the present application, the fault identification module 303 is further configured to: Starting from different segments of the M segments of each sample cable, the same k R And the resistance per unit length of the subsequent segments is corrected by the first correction rule to simulate short circuit faults at different locations; and / or, starting from the same segment of the M segments of each sample cable, different k R The resistance per unit length of the subsequent segment is corrected by the first correction rule to simulate short-circuit faults of different short-circuit degrees; Keep the capacitance, inductance and conductance per unit length of each section of each sample cable and the resistance per unit length of other sections constant; The capacitance per unit length of each sample cable is measured using different k C (k C The larger the value, the smaller the partial discharge fault) and the second correction rule is used to simulate partial discharge faults with different air gap sizes; The conductance per unit length of each sample cable is measured using different k σ (The greater the insulation aging, the greater the k σ The larger the value, the better) and corrected by the second correction rule to simulate insulation aging faults with different insulation aging degrees.

[0085] In another exemplary embodiment of the present application, the fault identification module 302 is further configured to: According to the following formula, starting from the Mth segment of each sample cable at each frequency point in each fault mode or normal mode, recursively calculate the input impedance of the first segment of each sample cable at each frequency point in each fault mode and normal mode from back to front: : ; ; ; in, The first frequency of each sample cable under any fault mode or normal mode i The input impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i The characteristic impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i+1 segment input impedance, The first frequency of each sample cable under any fault mode or normal mode i The propagation constant of the segment, is the length of each section of the sample cable, = l / N, l is the total length of the sample cable; 、 、 、 The corresponding value is the first frequency point of each sample cable under any fault mode or normal mode. i The resistance, inductance, conductance and capacitance of the segment, .

[0086] In the embodiment of the present application, the above formula is used to calculate When the terminal impedance Z of the sample cable is L =∞ (indicates that the far end of the sample cable is open).

[0087] In another exemplary embodiment of the present application, the fault location module 304 is further configured to: Determine the fault location of the target cable according to the following fault location model: ; in, The fault distance is the straight-line distance from the starting end of the target cable (the end where the electrical signal is injected) to the fault point. is the typical wave speed preset according to the cable type, It is the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable.

[0088] In the embodiment of the present application, since the resonance order n=1 is the strongest and clearest fundamental frequency mode, in order to improve positioning reliability and simplify calculation, the fault location is calculated according to the fundamental frequency mode (resonance order n=1).

[0089] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 5As shown. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, memory and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store cable local defect detection data. The I / O interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a cable local defect detection method is implemented.

[0090] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0091] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0092] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0093] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0094] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0095] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0096] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0097] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0098] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for detecting local defects in a cable, characterized in that: The cable local defect detection method comprises: Generating N electrical signals of different frequencies within a preset frequency range and inputting them into a target cable, testing the input impedance of the target cable under the electrical signals of the N different frequencies, and obtaining the broadband impedance of the target cable within the frequency range; According to the broadband impedance of the target cable, the impedance amplitude and impedance phase of the target cable at each preset frequency point are generated, and the impedance matrix X∈R is constructed. N×2 ; Inputting the impedance matrix X into the constructed fault classification model, identifying and outputting the category of the target cable through the fault classification model; wherein the category includes fault or normal; If the target cable is faulty, the resonant frequency with the largest amplitude in the broadband impedance amplitude spectrum of the target cable is extracted, and the fault location of the target cable is determined by constructing a fault location model based on the resonant frequency with the largest amplitude and the typical wave velocity of the target cable.

2. The cable local defect detection method according to claim 1, characterized in that: The fault classification model is constructed according to the following steps: Construct a training dataset: According to the constructed cable distribution parameter equivalent model, the transmission line equation is established; Determine an electrical parameter calculation formula at any frequency based on the transmission line equation, and calculate the electrical parameters per unit length of each sample cable at any frequency using the electrical parameter calculation formula; Correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode; Within a preset frequency range, N frequency points are set for each sample cable according to the preset frequency range and frequency scanning step, and the input impedance of the head end of each sample cable under each fault mode and each frequency point in the normal mode is calculated according to the electrical parameters per unit length of each sample cable under any frequency in the normal mode; Based on the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode, the impedance amplitude and impedance phase of the head end of each sample cable at each frequency point in each fault mode and normal mode are generated, and broadband impedance spectrum data of each sample cable at N frequency points in each fault mode and normal mode are obtained to form a training data set; According to the training data set, a CNN-based classification model is trained to obtain the fault classification model.

3. The cable local defect detection method according to claim 2, characterized in that: The correction rules include a first correction rule, a second correction rule and a third correction rule, wherein: The first amendment rule includes: R fault = R k R ,k R ∈[0.001,0.1]; The second revised rule includes: C fault = C k C ,k C ∈[0.2,0.5]; The third revised rule includes: G aging = G k σ ,k σ ∈[10,1000]; Among them, R fault is the corrected resistance per unit length, C fault is the corrected capacitance per unit length, G aging is the corrected conductance per unit length, R represents the resistance per unit length, G It represents the insulation conductivity per unit length, C Represents capacitance per unit length, k R is the resistance correction coefficient, k C is the capacitance correction factor, k σ is the conductance correction factor.

4. The cable local defect detection method according to claim 3, characterized in that: The failure modes include short circuit failure, partial discharge failure and insulation aging failure; The step of correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule to obtain the electrical parameter per unit length of each sample cable at any frequency of each fault mode specifically includes: Correcting the resistance per unit length of each of the sample cables using the first correction rule while maintaining the inductance, conductance, and capacitance per unit length of each of the sample cables unchanged, thereby obtaining electrical parameters per unit length of each of the sample cables at any frequency of a short-circuit fault; Correcting the capacitance per unit length of each of the sample cables using the second correction rule while keeping the inductance, conductance, and resistance per unit length of each of the sample cables unchanged, thereby obtaining electrical parameters per unit length of each of the sample cables at any frequency of a partial discharge fault; The conductance per unit length of each of the sample cables is corrected by the third correction rule, while the inductance, capacitance, and resistance per unit length of each of the sample cables are kept unchanged, thereby obtaining the electrical parameters per unit length of each of the sample cables at any frequency of insulation aging failure.

5. The cable local defect detection method according to claim 4, characterized in that: The step of constructing a training data set further includes: Before correcting at least one electrical parameter per unit length of each sample cable at any frequency according to a preset correction rule, each sample cable is divided into M segments from the head end to the terminal end of each sample cable; Correcting the resistance per unit length of each of the sample cables by using the first correction rule includes: Starting from different segments of the M segments of each sample cable, the same k R And the resistance per unit length of the subsequent segments is corrected by the first correction rule to simulate short circuit faults at different locations; and / or, starting from the same segment of the M segments of each sample cable, different k R and correcting the resistance per unit length of the subsequent segments using the first correction rule to simulate short-circuit faults of different short-circuit degrees; Maintaining the capacitance, inductance, and conductance per unit length of each section of each sample cable and the resistance per unit length of other sections unchanged; The second correction rule is used to correct the capacitance per unit length of each sample cable, including: The capacitance per unit length of each sample cable is measured using different k C And the correction is performed by the second correction rule to simulate partial discharge faults with different air gap sizes; The conductance per unit length of each sample cable is corrected by the third correction rule, including: The conductance per unit length of each sample cable is measured using different k σ The second correction rule is used to perform correction to simulate insulation aging faults with different insulation aging degrees.

6. The cable local defect detection method according to claim 2, characterized in that: Calculating the input impedance of the head end of each sample cable at each frequency point in each fault mode and normal mode according to the electrical parameters per unit length of each sample cable at any frequency in each fault mode and normal mode specifically includes: According to the following formula, starting from the Mth segment of each sample cable at each frequency point in each fault mode or normal mode, the impedance input impedance of the first segment of each sample cable at each frequency point in each fault mode and normal mode is calculated recursively from back to front. : ； ； ； in, The first frequency of each sample cable under any fault mode or normal mode i The input impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i The characteristic impedance of the segment, The first frequency of each sample cable under any fault mode or normal mode i +1 segment input impedance, The first frequency of each sample cable under any fault mode or normal mode i The propagation constant of the segment, is the length of each section of the sample cable, 、 、 、 Corresponding to the first frequency point of each sample cable in any fault mode or normal mode i The resistance, inductance, conductance and capacitance of the segment, , is the imaginary unit, represents the angular frequency, ω = 2 πf , f Indicates the frequency of any frequency point.

7. The cable local defect detection method according to claim 1, characterized in that: The fault classification model includes a CNN network and a fault classifier, where: The fault classifier is used to identify the category of the target cable according to the input overall feature vector, the category includes fault or normal, and the fault includes short circuit fault, partial discharge fault, and insulation aging fault; The CNN network includes: The first convolutional layer is used to extract local features of the impedance matrix X to obtain local frequency domain features; A maximum pooling layer is used to reduce the dimension of the local frequency domain features to obtain the local frequency domain features after dimensionality reduction; The second convolutional layer is used to extract deep features of the local frequency domain features after dimensionality reduction to obtain deep frequency domain features; The global average pooling layer is used to perform a global average pooling operation on the deep frequency domain features and output the overall feature vector.

8. The cable local defect detection method according to claim 7, characterized in that: The fault classifier comprises: A first fault classifier is used to identify the category of the target cable as a short circuit fault or normal according to the overall feature vector; A second fault classifier is used to identify the category of the target cable as a partial discharge fault or a normal condition based on the overall feature vector; The third classification model is used to identify the category of the target cable as insulation aging fault or normal according to the overall feature vector.

9. The cable local defect detection method according to claim 1, characterized in that: The fault location model includes: ； in, is the fault distance, is the typical wave speed preset according to the cable type, is the resonant frequency with the maximum amplitude in the broadband impedance amplitude spectrum of the target cable.

10. A cable local defect detection device, characterized in that: The cable local defect detection device comprises: An impedance testing module is configured to generate N electrical signals of different frequencies within a preset frequency range and input them into a target cable, thereby testing the input impedance of the target cable under the electrical signals of N different frequencies and obtaining the broadband impedance of the target cable within the frequency range; The impedance matrix construction module is used to generate the impedance amplitude and impedance phase of the target cable at each preset frequency point according to the broadband impedance of the target cable, and construct the impedance matrix X∈R N×2 ; A fault identification module is configured to input the impedance matrix X into the constructed fault classification model, identify and output the category of the target cable through the fault classification model; wherein the category includes fault or normal; The fault location module is used to extract the resonant frequency with the largest amplitude in the broadband impedance amplitude spectrum of the target cable if the target cable is faulty, and determine the fault location of the target cable through the constructed fault location model based on the resonant frequency with the largest amplitude and the typical wave velocity of the target cable.

Images