Cable fault identification method and device based on improved MVMD-KSD-PEPM, and medium

Abstract

The invention discloses a cable fault identification method and device based on improved MVMD-KSD-PEPM, and a medium, and the method comprises the steps: collecting a cable fault current x (t); decomposing the cable fault current through multivariate variational mode decomposition (MVMD) to obtain a plurality of intrinsic mode components (IMF); evaluating a similarity index between the IMF component and the fault current x (t) by using a KSD (Kollogov-Schmidunv Distance) to obtain a characteristic component; and calculating a permutation entropy partial mean value of the feature components, and performing fault identification by taking the permutation entropy partial mean value as a feature vector. The characteristics of different fault current signals of the cable can be better reflected, and the accuracy and effectiveness of cable fault identification are improved.

Description

Technical Field

[0001] The present invention belongs to the field of cable fault diagnosis, relates to a fault identification technology in a cable fault, and in particular to a cable fault identification method, a device and a medium based on an improved MVMD-KSD-PEPM. Background Art

[0002] With the rapid development of cities, underground cables have many advantages over overhead cables and are widely used. However, with the growth of social economy, the continuous increase in cable load and the barbaric construction on the ground, underground cables are prone to failure. In addition, due to the transformation of the power grid, the relocation and repair of cables, and the changes in the landform, the original drawings can no longer reflect the direction and depth of the cables, and the maintenance and management problems of underground cables are increasing. Because the underground cables span a large area, it is very difficult to find the fault point. If the cable fault cannot be repaired in time, it will not only waste a lot of manpower and material resources, but may also cause power outages. Therefore, it is of urgent practical significance to quickly and accurately locate the cable fault point.

[0003] Pre-location of cable faults requires extraction of weak useful signals under strong background noise and interference signals. Common noise reduction methods include wavelet transform, empirical mode decomposition, and adaptive noise complete set modal decomposition. Although wavelet transform can extract effective signals from measured signals, wavelet basis functions are difficult to select, and the noise reduction effects of different wavelet bases are inconsistent. Empirical mode decomposition and other methods are prone to problems such as modal aliasing and endpoint effects when decomposing signals. In contrast, the improved MVMD has a low probability of modal aliasing during the decomposition and transformation process, small instantaneous energy loss, and a low possibility of modal loss. It is not prone to problems such as modal aliasing and decomposition error accumulation, and is not affected by fault resistance.

[0004] In terms of fault identification, traditional cable fault diagnosis methods mainly include research methods based on signal detection and mathematical models. With the access of multi-energy loads and the interconnection of AC and DC at multiple voltage levels, the complexity of the power system continues to increase. Traditional cable fault diagnosis methods can no longer meet the requirements of modern power systems for stability and accuracy. With the gradual construction and advancement of smart grids and the rapid development of communication technology, the rapid and stable transmission of massive data provides a guarantee for the further training of artificial intelligence and also provides opportunities for data-driven fault diagnosis. Among them, support vector machine is an algorithm based on structural risk minimization through high-order operations and inner product operations. It is one of the classic classification algorithms in the field of machine learning and has good classification and recognition capabilities. Summary of the invention

[0005] The purpose of the present invention is to provide a cable fault identification method, device and medium based on the improved MVMD-KSD-PEPM, which can better reflect the characteristics of different fault current signals of the cable and improve the accuracy and effectiveness of cable fault identification.

[0006] In order to achieve the above object, the solution of the present invention is:

[0007] A cable fault identification method based on improved MVMD-KSD-PEPM, comprising:

[0008] Step A, collecting cable fault current x(t);

[0009] Step B, decomposing the cable fault current by multivariate variational mode decomposition (MVMD) to obtain multiple intrinsic mode components (IMF);

[0010] Step C, using the Kolmogorov-Smirnov distance KSD to evaluate the similarity index between the IMF component and the fault current x(t), and obtain the characteristic component;

[0011] Step D: Calculate the partial mean of the permutation entropy of the characteristic component and use it as a characteristic vector for fault identification.

[0012] In the above step B, the implementation process of multivariate variational mode decomposition MVMD is as follows:

[0013] Extract K predefined multivariate modulation oscillations u from the fault current x(t) k (t), its expression is:

[0014]

[0015] The cost function of MVMD is expressed as:

[0016]

[0017] In the formula, is the analytical modulation signal with channel number c and mode number k; is the partial derivative with respect to time; ω k is the center frequency of mode k;

[0018] The corresponding constraints are:

[0019]

[0020] in,

[0021]

[0022] Among them, x c(t) is the input data of the cth data channel, and C is the number of data channels included in the fault current x(t);

[0023] The corresponding augmented Lagrangian function is:

[0024]

[0025] Where λ is the Lagrangian operator and α is the penalty factor.

[0026] In the above step B, a comprehensive index S is constructed based on the correlation coefficient, approximate entropy and standard deviation, and the modal number K and the penalty factor α are optimized.

[0027]

[0028] In the formula, μ is a constant; r i is the correlation coefficient; h ae is the approximate entropy; i std is the standard deviation.

[0029] The above correlation coefficient is calculated according to the following formula:

[0030]

[0031] In the formula, r i is the correlation coefficient; y i are the eigenmode components; is the average value of N modal components, N is the total number of modal components; x is the fault current, is the mean fault current;

[0032] The algorithm of the approximate entropy is as follows:

[0033] The fault current x(t) is decomposed into R small sections from time 0 to time T, and the fault current of each small section is recorded as z u , where u=1,2,…,R; establish time series Z R ={z1,z2,…z R}, the time series Z R Reconstruct the m-dimensional phase space to obtain the sequence: Z m (i) = {z(s), z(s+1), …, z(s+m-1)}, s = 1, 2, …, R-m+1, m>2; the distance between the two dimensions is:

[0034] d=max|Z m (s)-Z m (j)|,j≠s,s≤j≤Rm

[0035] Where max is the maximum difference between the corresponding elements of the two-dimensional time series, and the similarity tolerance r=0.2*std is defined, where std is Z N Standard deviation; the number of vectors whose distance is less than the similarity tolerance r is the ratio of num to the total vector:

[0036]

[0037] Take the logarithm of the above ratio and find the average value of each s where m is a vector. The formula is:

[0038]

[0039] Change the phase space reconstruction dimension m = m + 1, repeat the above steps to obtain The approximate entropy of the real cable phase current is calculated as:

[0040]

[0041] The standard deviation is calculated according to the following formula,

[0042]

[0043] In the formula, x t is the fault current at time t, is the mean fault current.

[0044] The specific process of the above step C is:

[0045] First, calculate the empirical distribution function E x (·):

[0046]

[0047] In the formula, x (i) is a simple random sample of the fault current x(t), i = 1, 2, ..., f, and the components are arranged in descending order as follows: (1) <x (2) <… <x (n) , where x (i) The frequency of occurrence is n i , n1+n2+…+n f =n;

[0048] Define the KSD between each IMF component and the fault current as D, and its calculation formula is:

[0049]

[0050] Among them, F n (x) is the empirical distribution function of the IMF component, R n(x) is the empirical distribution function of the fault current;

[0051] The root mean square value of KSD is used as the threshold, and the IMF component with the most characteristic information is selected as the characteristic component based on this threshold; the threshold calculation formula is:

[0052]

[0053] In the above step D, the calculation process of the partial mean of permutation entropy is as follows:

[0054] Step D1, calculate the permutation entropy:

[0055] Decompose the fault current x(t) into n equal parts from 0 to T t The fault current of each small section is recorded as where w = 1, 2, ..., n t ; Consider a time series dataset n t is the number of real samples, and the permutation entropy of each sample is calculated with a sliding window of length L, and reconstructed within the sliding window The phase space of is described as:

[0056]

[0057] Where m is the embedding dimension, τ is the time delay, and M = L-(m-1)τ is the number of vectors in the reconstructed image space. The j-th row of the reconstruction matrix is ​​arranged in ascending order to form a new sequence, which is expressed as:

[0058]

[0059] Where a1 represents the index of the minimum value in the jth row of the reconstruction matrix, a m represents the index of the maximum value in the jth row of the reconstruction matrix; the sequence thus composed has m! permutation possibilities, and the permutation entropy of the fault current is expressed as:

[0060]

[0061] Where P a represents the probability distribution of symbols, K≤m!;

[0062] Step D2, calculate the skewness PE of the permutation entropy k :

[0063] PE k =3(PE α -PE m ) / PE s

[0064] Among them, PE αis the average value of PE; PE m is the median of PE; PE s is the standard deviation of PE;

[0065] Step D3, calculate the partial mean of the permutation entropy:

[0066] PEPM=(1+|PE k | / 3)*PE α .

[0067] In the above step D, the feature vector is imported as an input layer variable into the support vector machine for fault identification. The optimal hyperplane classification function of the support vector machine is:

[0068]

[0069] In the formula, μ d is a Lagrange multiplier, and 0≤μ d ≤ψ, ψ is the penalty factor; K(x i ·x j ) is the kernel function, x i ·x j is the variable of the kernel function; D is the number of points to be classified; γ is the classification threshold determined by the training sample.

[0070] A computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor; when the processor executes the computer program, the steps of the cable fault identification method based on the improved MVMD-KSD-PEPM as described above are implemented.

[0071] A computer-readable storage medium stores a computer program; when the computer program is executed by a processor, the steps of the cable fault identification method based on the improved MVMD-KSD-PEPM as described above are implemented.

[0072] After adopting the above scheme, the beneficial effect of the present invention is that it provides a solution to the difficulty of obtaining characteristic vectors in cable fault identification. First, the improved MVMD is used to decompose the detection signal to obtain the intrinsic mode component, which can better retain the time-varying characteristics of the detection signal and is not easy to produce modal aliasing and decomposition errors; secondly, KSD is used as an indicator to select characteristic components, and a threshold is constructed to obtain IMF components with more characteristic information, which can significantly improve the fault identification accuracy; finally, the permutation entropy partial mean, a comprehensive indicator based on permutation entropy and skewness, can better reflect the characteristics of different fault current signals of the cable and improve the accuracy and effectiveness of cable fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1is a flow chart of the present invention;

[0074] Figure 2 It is a flowchart of improving MVMD;

[0075] Figure 3 It is the principle diagram of support vector machine. DETAILED DESCRIPTION

[0076] The technical solutions and beneficial effects of the present invention will be described in detail below with reference to the accompanying drawings.

[0077] The embodiment of the present invention takes a phase A grounding fault as an example. Figure 1 A flow chart of a cable fault identification method based on improved MVMD-KSD-PEPM is given. The method includes the following steps:

[0078] Step 1: Data collection. Collect the cable A phase fault current I a , denoted as x(t);

[0079] Step 2: Multivariate variational mode decomposition (MVMD). Decompose the fault current of phase A through MVMD to obtain the intrinsic mode component IMF;

[0080] From the input data x(t) containing C data channels, K predefined multivariate modulated oscillators u are extracted. k (t), its expression is:

[0081]

[0082] The cost function of MVMD can be expressed as:

[0083]

[0084] In the formula, is the analytical modulation signal with channel number c and mode number k; is the partial derivative with respect to time; ω k is the center frequency of mode k.

[0085] The corresponding constraints are:

[0086]

[0087] in,

[0088]

[0089] Among them, x c (t) is the input data of the cth data channel;

[0090] The corresponding augmented Lagrangian function is:

[0091]

[0092] Where λ is the Lagrangian operator and α is the penalty factor.

[0093] Step 3: Parameter optimization. Based on the correlation coefficient, approximate entropy and standard deviation, a comprehensive index S is constructed to optimize the modal number K and penalty factor α of MVMD;

[0094] First, calculate the correlation coefficient, approximate entropy, and standard deviation.

[0095] Correlation coefficient:

[0096]

[0097] In the formula, r i is the correlation coefficient; y i are the eigenmode components; is the average value of N modal components, N is the total number of modal components; x is the fault current I a , is the mean fault current.

[0098] Approximate entropy:

[0099] Approximate entropy can describe the change pattern and unpredictability of time series data. The approximate entropy algorithm is as follows:

[0100] The fault current x(t) is decomposed into R small sections from time 0 to time T, and the fault current of each small section is recorded as z u , where u=1,2,…,R. Establish time series Z R ={z1,z2,…z R}, the time series Z R Reconstructing the m (m>2) dimensional phase space gives the sequence: Z m (i) = {z(s), z(s+1), …, z(s+m-1)}, s = 1, 2, …, R-m+1. The distance between the two dimensions is:

[0101] d=max|Z m (s)-Z m (j)|,j≠s,s≤j≤Rm

[0102] Where max is the maximum difference between the corresponding elements of the two-dimensional time series, and the similarity tolerance r=0.2*std is defined, where std is Z R The number of vectors whose distance is less than the similarity tolerance r is the ratio of num to the total vector:

[0103]

[0104] Take the logarithm of the above ratio and find the average value of each s where m is a vector. The formula is:

[0105]

[0106] Change the phase space reconstruction dimension m = m + 1, repeat the above steps to obtain The approximate entropy of the real cable phase current can be obtained by calculation:

[0107]

[0108] Standard Deviation:

[0109]

[0110] In the formula, x t is the fault current at time t, is the mean fault current.

[0111] Constructing the comprehensive indicator S:

[0112]

[0113] Where μ is a constant.

[0114] Here, by obtaining the changes in the comprehensive indicators under different K and α values, when the comprehensive indicator S is the largest, the correlation coefficient is small, the modes are relatively independent, and the probability of modal aliasing is minimized, so as to obtain a better decomposition effect.

[0115] Step 4: Preliminary noise reduction. Use the optimized MVMD to decompose the cable fault current to reduce modal aliasing;

[0116] Figure 2 The flowchart of improved MVMD is given.

[0117] Step 5: Feature component selection. The Kolmogorov-Smirnov distance (KSD) is used to evaluate the similarity index between the IMF component and the original signal to select the feature component;

[0118] First, calculate the empirical distribution function E x (·):

[0119]

[0120] Where: x (i) (i=1, 2, ..., f) is a simple random sample of the fault current x(t), and the components are arranged in descending order as follows:(1) <x (2) <… <x (n) Among them, x (i) The frequency of occurrence is n i (n1+n2+…+n f =n)

[0121] The KSD between each IMF component and the original current signal is defined as D, and its calculation formula is:

[0122]

[0123] Among them, F n (x) is the empirical distribution function of the IMF component, R n (x) is the empirical distribution function of the original current signal.

[0124] The root mean square value of KSD is used as the threshold, and the IMF component with more characteristic information is selected as the characteristic component based on this threshold. The threshold calculation formula is:

[0125]

[0126] Step 6: Construct feature vector. Based on permutation entropy (PE) and skewness, a comprehensive index, partial mean of permutation entropy (PEPM), is constructed to reflect the characteristics of different cable line faults, thereby constructing feature vectors and improving fault identification accuracy.

[0127] The calculation process of the partial mean of permutation entropy is:

[0128] (1) Calculate the permutation entropy:

[0129] Permutation Entropy (PE) is a dynamic anomaly detection method for time series. Its uniqueness lies in that it can accurately present the characteristic differences between normal points and fault points by amplifying small changes in the signal. Therefore, permutation entropy has excellent ability to capture abnormal situations in time series and can effectively distinguish normal data from faulty data. Its calculation method is as follows:

[0130] Decompose the fault current x(t) into n equal parts from 0 to T t The fault current of each small section is recorded as where w = 1, 2, ..., n t Consider a time series dataset n t is the number of real samples, and the permutation entropy of each sample is calculated with a sliding window of length L, and reconstructed within the sliding window The phase space of can be described as:

[0131]

[0132] Where m is the embedding dimension, τ is the time delay, and M = L-(m-1)τ is the number of vectors in the reconstructed image space. Arrange the jth row of the reconstruction matrix in ascending order to form a new sequence, which can be expressed as:

[0133]

[0134] Where a1 represents the index of the minimum value in the jth row of the reconstruction matrix, a m represents the index of the maximum value in the jth row of the reconstruction matrix. The sequence thus formed consists of m! permutation possibilities. The permutation entropy of the real cable phase current can be expressed as:

[0135]

[0136] Where P a represents the probability distribution of symbols, K≤m!.

[0137] (2) Calculate the skewness PE of the permutation entropy k :

[0138] PE k =3(PE α -PE m ) / PE s

[0139] Among them, PE α is the average value of PE; PE m is the median of PE; PE s is the standard deviation of PE.

[0140] (3) Calculate the partial mean of permutation entropy:

[0141] PEPM=(1+|PE k | / 3)*PE α

[0142] Step 7: Fault identification. The feature vector obtained in step 6 is imported as the input layer variable into the support vector machine for fault identification.

[0143] The optimal hyperplane classification function of the support vector machine is:

[0144]

[0145] In the formula, μ d is a Lagrange multiplier, and 0≤μ d ≤ψ, ψ is the penalty factor; K(x i ·xj ) is the kernel function, x i ·x j is the variable of the kernel function; D is the number of points to be classified; γ is the classification threshold determined by the training sample.

[0146] Figure 3 A principle diagram of a support vector machine is given. In the diagram, H is a hyperplane, H1 and H2 are classification surfaces that are parallel to H and have equal distances, and Margin is the distance between H1 and H2. When the distance is the largest, H is the optimal hyperplane.

[0147] An embodiment of the present invention further provides another computer device, including a processor and a memory configured to store a computer program that can be run on the processor; wherein, when the processor is configured to run the computer program, the method steps in the aforementioned embodiment are executed.

[0148] In practical applications, the processor includes a field programmable gate array (FPGA), and the processor may be a central processing unit (CPU) or a digital signal processor (DSP). It is understandable that for different devices, the electronic device used to implement the function of the processor may be other, and the embodiment of the present invention does not specifically limit it.

[0149] The above-mentioned memory can be a volatile memory (volatile memory), such as a random access memory (RAM); or a non-volatile memory (non-volatile memory), such as a read-only memory (ROM), a flash memory, a hard disk (HDD) or a solid-state drive (SSD); or a combination of the above-mentioned types of memory, and provide instructions and data to the processor.

[0150] In an exemplary embodiment, an embodiment of the present invention further provides a computer-readable storage medium for storing a computer program.

[0151] Optionally, the computer-readable storage medium can be applied to any one of the methods in the embodiments of the present invention, and the computer program enables the computer to execute the corresponding processes implemented by the processor in each method in the embodiments of the present invention. For the sake of brevity, they are not described here.

[0152] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0153] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0154] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0155] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0156] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0157] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0158] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A cable fault identification method based on improved MVMD-KSD-PEPM, characterized in that: include, Step A, collecting cable fault current x(t); Step B, decomposing the cable fault current by multivariate variational mode decomposition (MVMD) to obtain multiple intrinsic mode components (IMF); Step C, using the Kolmogorov-Smirnov distance KSD to evaluate the similarity index between the IMF component and the fault current x(t), and obtain the characteristic component; Step D: Calculate the partial mean of the permutation entropy of the characteristic component and use it as a characteristic vector for fault identification.

2. The method according to claim 1, characterized in that: In step B, the implementation process of multivariate variational mode decomposition MVMD is as follows: Extract K predefined multivariate modulation oscillations u from the fault current x(t) k (t), its expression is: The cost function of MVMD is expressed as: In the formula, is the analytical modulation signal with channel number c and mode number k; is the partial derivative with respect to time; ω k is the center frequency of mode k; The corresponding constraints are: in, Among them, x c (t) is the input data of the cth data channel, and C is the number of data channels included in the fault current x(t); The corresponding augmented Lagrangian function is: Where λ is the Lagrangian operator and α is the penalty factor.

3. The method according to claim 2, characterized in that: The step B also includes constructing a comprehensive index S based on the correlation coefficient, approximate entropy and standard deviation, optimizing the modal number K and the penalty factor α, In the formula, μ is a constant; r i is the correlation coefficient; h ae is the approximate entropy; i std is the standard deviation.

4. The method according to claim 3, characterized in that: The correlation coefficient is calculated according to the following formula: In the formula, r i is the correlation coefficient; y i are the eigenmode components; is the average value of N modal components, N is the total number of modal components; x is the fault current, is the mean fault current; The algorithm of the approximate entropy is as follows: The fault current x(t) is decomposed into R small sections from time 0 to time T, and the fault current of each small section is recorded as z u , where u=1,2,…,R; establish time series Z R ={z1,z2,…z R }, the time series Z R Reconstruct the m-dimensional phase space to obtain the sequence: Z m (i) = {z(s), z(s+1), …, z(s+m-1)}, s = 1, 2, …, R-m+1, m>2; the distance between the two dimensions is: d=max|Z m (s)-Z m (j)|,j≠s,s≤j≤R-m Where max is the maximum difference between the corresponding elements of the two-dimensional time series, and the similarity tolerance r=0.2*std is defined, where std is Z N Standard deviation; the number of vectors whose distance is less than the similarity tolerance r is the ratio of num to the total vector: Take the logarithm of the above ratio and find the average value of each s where m is a vector. The formula is: Change the phase space reconstruction dimension m = m + 1, repeat the above steps to obtain The approximate entropy of the real cable phase current is calculated as: The standard deviation is calculated according to the following formula, In the formula, x t is the fault current at time t, is the mean fault current.

5. The method according to claim 1, characterized in that: The specific process of step C is: First, calculate the empirical distribution function E x (·): In the formula, x (i) is a simple random sample of the fault current x(t), i = 1, 2, ..., f, and the components are arranged in descending order as follows: (1) <x (2) <… <x (n) , where x (i) The frequency of occurrence is n i , n1+n2+…+n f =n; Define the KSD between each IMF component and the fault current as D, and its calculation formula is: Among them, F n (x) is the empirical distribution function of the IMF component, R n (x) is the empirical distribution function of the fault current; The root mean square value of KSD is used as the threshold, and the IMF component with the most characteristic information is selected as the characteristic component based on this threshold; the threshold calculation formula is:

6. The method according to claim 1, characterized in that: In step D, the calculation process of the partial mean of the permutation entropy is as follows: Step D1, calculate the permutation entropy: Decompose the fault current x(t) into n equal parts from 0 to T t The fault current of each small section is recorded as where w = 1, 2, ..., n t ; Consider a time series dataset n t is the number of real samples, and the permutation entropy of each sample is calculated with a sliding window of length L, and reconstructed within the sliding window The phase space of is described as: Where m is the embedding dimension, τ is the time delay, and M = L-(m-1)τ is the number of vectors in the reconstructed image space. The j-th row of the reconstruction matrix is ​​arranged in ascending order to form a new sequence, which is expressed as: Where a1 represents the index of the minimum value in the jth row of the reconstruction matrix, a m represents the index of the maximum value in the jth row of the reconstruction matrix; the sequence thus composed has m! permutation possibilities, and the permutation entropy of the fault current is expressed as: Where P a represents the probability distribution of symbols, K≤m!; Step D2, calculate the skewness PE of the permutation entropy k : ON k =3(PE α -ON m ) / ON s Among them, PE α is the average value of PE; PE m is the median of PE; PE s is the standard deviation of PE; Step D3, calculate the partial mean of the permutation entropy: PEPM=(1+|PE k | / 3)*PE α 。 7. The method according to claim 1, characterized in that: In step D, the feature vector is imported as an input layer variable into a support vector machine for fault identification. The optimal hyperplane classification function of the support vector machine is: In the formula, μ d is a Lagrange multiplier, and 0≤μ d ≤ψ, ψ is the penalty factor; K(x i ·x j ) is the kernel function, x i ·x j is the variable of the kernel function; D is the number of points to be classified; γ is the classification threshold determined by the training sample.

8. A computer 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 computer program, the steps of the cable fault identification method based on the improved MVMD-KSD-PEPM as described in any one of claims 1 to 7 are implemented.

9. A computer-readable storage medium storing a computer program; characterized in that: When the computer program is executed by a processor, the steps of the cable fault identification method based on the improved MVMD-KSD-PEPM as described in any one of claims 1 to 7 are implemented.

Images