Disconnecting switch fault diagnosis method and system based on Levy-FOX-VMD-SVM

Abstract

The invention discloses an isolating switch mechanical structure fault diagnosis method based on Levy-FOX-VMD and SVM. Firstly, a fault simulation experiment is carried out on a certain type of disconnecting switch, and a vibration signal is measured; then, the Levy-FOX is adopted to optimize the VMD algorithm, and optimal parameters alpha and k are found out; secondly, decomposing a non-stationary vibration signal into a plurality of intrinsic mode functions (IMF) by adopting VMD, and calculating a plurality of entropies of each layer of IMF to form a multivariate feature fusion matrix; and carrying out dimensionality reduction on the multivariate entropy matrix by using a PCA algorithm. And finally, inputting the feature matrix into an SVM for training to realize fault diagnosis of the mechanical structure of the disconnecting switch.

Description

Technical Field

[0001] The invention belongs to the field of disconnector mechanical structure fault diagnosis, and in particular relates to a disconnector fault diagnosis method and system based on Levy-FOX-VMD-SVM. Background Art

[0002] At present, China's power industry is ushering in an unprecedented period of development and prosperity. The rapid development of the core information technology of the intelligent Internet of Things is gradually reshaping the traditional face of the power industry. This technological wave has led to the emergence of a large number of new technologies, new products and new business models, providing a strong impetus for the construction of smart grids. With the continuous advancement of smart Internet of Things technology, the requirements for the operation efficiency and safety of power equipment are becoming higher and higher. Especially in the power grid, the operation status of high-voltage disconnectors has received great attention. In the past decade, with the rapid expansion of power grid infrastructure, the establishment of substations and the number of disconnectors have increased significantly. The mechanical operating environment of disconnectors is complex and changeable, such as being attacked by wind, rain and sun. In addition, the operation frequency of disconnectors is not high, which can easily lead to rust, stagnation, inadequate closing or even refusal to move in the transmission part. In severe cases, it may also cause the transmission mechanism to break or the line to trip, so it is particularly important to monitor the status and perform real-time maintenance of disconnectors.

[0003] When the high-voltage disconnector is not in place, the mechanism is stuck, loose, or the three phases are out of phase, various transmission components start and stop, which causes equipment vibration. The vibration frequency of electrical equipment is single when it is running, and the corresponding vibration signal characteristics are also unique. Modal analysis is usually used to analyze them. If the mechanical structure of the disconnector changes or fails, its frequency and modal characteristics of the vibration signal will also change. Therefore, this method can be used to diagnose whether the disconnector is in a normal state or a faulty state.

[0004] When existing studies use EMD decomposition, we need to use the maximum and minimum values ​​of the signal to calculate the envelope, and the extreme values ​​of the signal are likely to be interfered by noise, so EMD is sensitive to noise; EMD also has the problem of endpoint effect, that is, at the endpoint of the signal, due to the lack of information about the previous and next frames, the envelope surface obtained may be inaccurate, and in the later decomposition process, inaccurate IMF components may be obtained. The method of using a single VMD to decompose the signal is still susceptible to noise interference when processing noisy signals, which will also lead to inaccurate decomposition results. Summary of the invention

[0005] In order to solve the deficiencies in the prior art, the present invention provides a disconnector fault diagnosis method and system based on Levy-FOX-VMD-SVM, which uses the Fox (FOX) algorithm optimized by Levy flight strategy to decompose the collected vibration signal, obtain the corresponding intrinsic mode sub-signal and calculate the various entropies of each layer of energy to form a feature matrix. Then, the obtained feature matrix is ​​processed by dimensionality reduction using PCA (Principal Component Analysis). Finally, the feature vector is input into SVM (Support Vector Machine) for training to realize the fault diagnosis of the mechanical structure of the disconnector.

[0006] The present invention adopts the following technical solution.

[0007] The present invention provides a disconnector fault diagnosis method based on Levy-FOX-VMD-SVM, which is characterized by comprising:

[0008] The vibration signals of normal operation, inadequate opening and closing, three-phase out of phase and stuck state are extracted from the historical data of the disconnector. The four groups of vibration signals are calculated by VMD algorithm, and the vibration signals are decomposed into k layers of IMF signals. The minimum power spectrum entropy of each layer of IMF signals of the four groups of signals is calculated as the fitness function, and the Levy-FOX algorithm is used to optimize the parameters of VMD. The first group of optimal parameter values ​​obtained are used as the parameters of VMD. The vibration signals are re-divided into 200 groups for VMD, and the characteristic entropy values ​​of each layer of IMF signals are calculated to form a multivariate feature fusion matrix. The PCA algorithm is used to reduce the dimension of the multivariate feature fusion matrix and select the principal components to obtain a three-dimensional reconstruction space. The fault detection module is built using the SVM classifier, and the vibration signal of the mechanical structure of the disconnector to be tested is input. The fault detection module performs eigenvalue matching and outputs the fault detection result.

[0009] Preferably, the calculating the four groups of vibration signals by using a VMD algorithm to decompose the vibration signals into k layers of IMF signals comprises:

[0010] The IMF signal decomposed by VMD is:

[0011] u k (t) = A k (t)cos[φ k (t)]

[0012] Where: A k (t) is the IMF signal amplitude, φ k (t) is the instantaneous phase;

[0013] The VMD constraint requires that the sum of the bandwidths of each IMF center frequency be minimized.

[0014]

[0015] {u k} represents the IMF signal set of each layer; {ω k} represents the center frequency set of each layer of IMF; * is the convolution operation; is the gradient operation; δ(t) is the Dirac function; j represents the imaginary unit; f(t) represents the IMF signal of each layer as a function of time, α is the penalty factor; L is the Lagrangian operator, λ is the Lagrangian multiplier; λ(t) is the function of the Lagrangian multiplier with respect to time.

[0016] Preferably, the minimum power spectrum entropy of each layer of IMF signal is used as the fitness function to adaptively determine the optimal value of the number of IMF layers k and the penalty factor α.

[0017] Preferably, the minimum power spectrum entropy is calculated as follows:

[0018] S N (f) = |X F (f)| 2 / N

[0019]

[0020]

[0021] Where: f is the original signal, the image signal X represents the IMF components of each layer after VMD decomposition, |X F (f) | represents the Fourier transform of the image signal X, H X is the power spectrum entropy of the graph, ξ i is the percentage of the i-th spectrum power spectrum in the entire spectrum power spectrum, N is the number of sampling points of the IMF signal in each layer.

[0022] Preferably, the Levy-FOX algorithm is used to optimize the parameters of VMD, including:

[0023] Step 4.1, set the number of fox groups N = 50, the initial fox search step length is fixed to d = 100, the maximum number of iterations is set to R = 50, and the random number r∈[0,1],

[0024] The Levy flight function is used to optimize the FOX algorithm. The Levy flight function processes the search step length β of the FOX algorithm as follows to make it conform to the Levy distribution; u and υ are normal random variables;

[0025]

[0026]

[0027] β = 1.5;

[0028] Step 4.2, set the number of VMD decomposition levels k to range from 3 to 10, and optimize the penalty factor α from 1 to 5000;

[0029] Step 4.3, obtain a random number r, and the fox population starts to perform jumping foraging or random walking foraging. If r ≥ 0.5, jump foraging is performed; if r ≤ 0.5, random walking foraging is performed;

[0030] Step 4.4, when the fox finds a position with a smaller power spectrum entropy, update the optimal value. If the maximum number of iterations has not been reached, repeat step 4.3. After the maximum number of iterations is reached, output the optimal values ​​of α and k as the first set of optimal parameter values. The fox position update formula is as follows:

[0031]

[0032] Where: Jump is the jumping height of each fox in the algorithm, c and γ are random numbers between [0,1], and X it+1 For the latest location of the fox, X (it) The position of the fox before the update.

[0033] Preferably, the characteristic entropy values ​​include: approximate entropy, sample entropy, fuzzy entropy, Kolmogorov entropy, permutation entropy, modified conditional entropy, distribution entropy, spectral entropy, dispersion entropy, symbolic dynamic entropy, phase entropy, bubble entropy, grid distribution entropy, attention entropy and range entropy.

[0034] Preferably, the fault detection part includes building a fault detection module using an SVM classifier, inputting the characteristic values ​​of the four faults in the PCA three-dimensional reconstruction space into the fault detection module; inputting the vibration signal of the mechanical structure of the isolating switch to be tested, and the fault detection module performs characteristic value matching to output the fault detection result.

[0035] The present invention also provides a disconnector fault diagnosis system based on Levy-FOX-VMD-SVM, the system being a system used by the aforementioned disconnector fault diagnosis method based on Levy-FOX-VMD-SVM, comprising:

[0036] A data acquisition module is used to obtain historical vibration signal data of the disconnector;

[0037] VMD algorithm module, used to decompose vibration signals into IMF signals;

[0038] Levy-FOX algorithm module, used to optimize parameters of VMD;

[0039] The SVM algorithm module is used to diagnose the faults of the mechanical structure of the disconnector.

[0040] The present invention also provides a terminal, including a processor and a storage medium;

[0041] The storage medium is used to store instructions;

[0042] The processor is configured to operate according to the instructions to execute the steps according to the aforementioned method.

[0043] The present invention also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented.

[0044] The beneficial effect of the present invention is that, compared with the prior art, the present invention proposes a disconnector mechanical structure fault diagnosis method based on Levy-FOX-VMD and SVM, and the results show that the Levy-FOX algorithm has better convergence than the FOX algorithm and the PSO algorithm under the cec2022-F1 test function, and the SVM algorithm has a faster running rate than the single SVM algorithm. The proposed method is suitable for disconnector mechanical structure fault diagnosis, improves the iteration speed and convergence, and has a higher fault identification accuracy for disconnectors. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flow chart of the disconnector fault diagnosis method based on Levy-FOX-VMD-SVM in the present invention;

[0046] Figure 2 A schematic diagram of the original signal and the decomposed signal of the designated signal VMD decomposition in the present invention;

[0047] Figure 3 It is a schematic diagram of the fitness curve comparison of the FOX algorithm, Levy-FOX algorithm and PSO algorithm under the cec2022-F1 test function in the present invention;

[0048] Figure 4 It is a three-dimensional schematic diagram of the cec2022-F1 test function in the present invention;

[0049] Figure 5 It is a scatter diagram schematic diagram of the PCA algorithm in the present invention reducing the dimension of the multivariate feature entropy fusion matrix to three dimensions. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0051] like Figure 1 As shown, embodiment 1 of the present invention provides a disconnector fault diagnosis method based on Levy-FOX-VMD-SVM, comprising:

[0052] Step 1, extracting vibration signals of normal operation, inadequate opening and closing, three-phase sync and stuck state from historical data of the disconnector;

[0053] Step 2, such as Figure 2 As shown, the four groups of vibration signals are calculated using the VMD algorithm, and the vibration signals are decomposed into k-layer IMF signals;

[0054] Furthermore, the first 2000 data of the four groups of vibration signals collected are selected to form a 4×2000 matrix and input into the VMD algorithm. During the solution process, VMD decomposes the original vibration signal into k layers of IMFs and adaptively updates the optimal center frequency and bandwidth of each IMF. The specific steps of VMD are as follows:

[0055] The IMF decomposed by VMD can be expressed as:

[0056] u k (t) = A k (t)cos[φ k (t)] (1)

[0057] Where: A k (t) is the IMF signal amplitude; φ k (t) is the instantaneous phase.

[0058] The goal of VMD is to decompose a real-valued input signal into a discrete number of sub-signals {u k}, these sub-signals have a specific sparsity in reproducing the input. The sparsity of each pattern is its bandwidth in the spectral domain.

[0059] The bandwidth is estimated using the square norm of the gradient of the demodulated signal. The resulting constrained variational problem is as follows: The VMD constraint requires that the sum of the bandwidths of each IMF center frequency be minimized, so a VMD constrained variational problem is constructed to find the minimum bandwidth:

[0060]

[0061] st∑ k u k =f(t) (3)

[0062] Where: {u k} represents the IMF signal set of each layer; {ω k} represents the center frequency set of each layer of IMF; * is the convolution operation; is the gradient operation; δ(t) is the Dirac function; j represents the imaginary unit; f(t) represents the IMF signal of each layer represented by a function about time.

[0063] The augmented Lagrangian function is introduced to transform equation (2) into an unconstrained problem to solve the above constrained variation problem:

[0064]

[0065] Where: α is the penalty factor; L is the Lagrangian operator; <> is the inner product operation; λ is the Lagrangian multiplier; λ(t) is the function of the Lagrangian multiplier with respect to time.

[0066] Step 3, calculate the minimum power spectrum entropy of the IMF signals of each layer of the four groups of signals as the fitness function;

[0067] The fitness function can be used to adaptively determine the optimal number of IMFs k and penalty factor α to balance the number of IMFs and decomposition quality. The minimum power spectrum entropy of the IMF signal is extracted using Fourier transform:

[0068] S N (f) = |X F (f)| 2 / N (5)

[0069]

[0070]

[0071] Where: f is the original signal, the image signal X represents the IMF components of each layer after VMD decomposition, |X F (f) | represents the Fourier transform of the image signal X, H X is the power spectrum entropy of the graph, ξ i is the percentage of the i-th spectrum power spectrum in the entire spectrum power spectrum, that is, N is the number of sampling points of the IMF signal in each layer.

[0072] Step 4, such as Figure 3 , 4 As shown, the Levy-FOX algorithm is used to optimize the parameters of VMD, and the minimum power spectrum entropy is used as the fitness function for optimization;

[0073] Step 4.1, set the number of fox groups N = 50, the initial fox search step length is fixed to d = 100, the maximum number of iterations is set to R = 50, the random number r∈[0,1], and the Levy flight function is used to optimize the FOX algorithm. The Levy flight function processes the FOX algorithm search step length β as follows to make it conform to the Levy distribution. u and c are normal random variables;

[0074]

[0075] The three parameter values ​​are generally calculated as follows:

[0076]

[0077] β = 1.5;

[0078] Step 4.2, set the number of VMD decomposition levels k to range from 3 to 10, and optimize the penalty factor α from 1 to 5000;

[0079] Step 4.3, obtain a random number r, and the fox population starts to perform jumping foraging or random walking foraging. If r ≥ 0.5, jump foraging is performed; if r ≤ 0.5, random walking foraging is performed;

[0080] Step 4.4, when the fox finds a position with a smaller power spectrum entropy, update the optimal value. If the maximum number of iterations has not been reached, repeat step 4.3. After the maximum number of iterations is reached, output the optimal values ​​of α and k. The fox position update formula is as follows:

[0081]

[0082] Where: Jump is the jumping height of each fox in the algorithm, c and γ are random numbers between [0,1], and X it+1 For the latest location of the fox, X (it) The position of the fox before the update.

[0083] Step 5, using the optimal parameter values ​​obtained in step 4 as the parameters of VMD, dividing the vibration signal into 200 groups for VMD, and calculating the approximate entropy, sample entropy, fuzzy entropy, Kolmogorov entropy, permutation entropy, modified conditional entropy, distribution entropy, spectral entropy, dispersion entropy, symbolic dynamic entropy, phase entropy, bubble entropy, grid distribution entropy, attention entropy and range entropy of each layer of IMF of these 200 groups of signals to form a multivariate feature fusion matrix.

[0084] Furthermore, the number of vibration signal groups can be adjusted to any value according to actual conditions.

[0085] Step 6, such as Figure 5As shown in the figure, the PCA algorithm is used to reduce the dimension of the multivariate feature fusion matrix of each layer of IMF components, and the principal component selection is performed without losing the original information. The PCA solution steps are as follows:

[0086] A data set containing 200 sets of vibration signal samples and 15 features (types of entropy in the feature matrix): X = {x1, x2, ..., x 200}, reduce its dimension to m dimensions. Let the sample set be matrix X:

[0087]

[0088] Each row represents a sample, each column represents a feature, and the column number represents the dimension of the feature, a total of 15 dimensions. Output the sample set after dimensionality reduction: Y = {y1, y2, ..., y 200}.

[0089] Decentralize the matrix to get a new matrix X, that is, zero-mean each column, that is, subtract the mean of this column

[0090]

[0091] The required matrix X is still a 200×15 matrix:

[0092] Compute the covariance matrix of the decentralized matrix X:

[0093]

[0094] The matrix is ​​a 15×15 matrix.

[0095] Perform eigendecomposition on the covariance matrix C to find the eigenvalues ​​of the covariance matrix and the corresponding eigenvectors:

[0096] The eigenvectors are arranged in descending order from left to right according to the corresponding eigenvalues, and the first m columns are taken to form the matrix W, that is, a 15×m-order matrix. The sample features after dimension reduction to m dimensions are calculated through Y=XW, that is, a 200×m-order matrix.

[0097] After testing: when m≥1, the m principal components already contain the main information of 15 feature entropies (the total variance contribution rate exceeds 95%). In order to achieve the effect of dimensionality reduction, m=1 is taken, that is, the multivariate feature entropy matrix is ​​reduced in dimension to generate a three-dimensional PCA reconstruction space.

[0098] Step 7, using the SVM classifier to build a fault detection module, and inputting the characteristic values ​​of the four faults in the PCA three-dimensional reconstruction space into the fault detection module;

[0099] Step 8: Input the vibration signal of the mechanical structure of the isolating switch to be tested, and the fault detection module performs feature value matching to output the fault detection result.

[0100] Embodiment 2 of the present invention provides a disconnector fault diagnosis system based on Levy-FOX-VMD-SVM, the system being a system used by the aforementioned disconnector fault diagnosis method based on Levy-FOX-VMD-SVM, and characterized in that it comprises:

[0101] A data acquisition module is used to obtain historical vibration signal data of the disconnector;

[0102] VMD algorithm module, used to decompose vibration signals into IMF signals;

[0103] Levy-FOX algorithm module, used to optimize parameters of VMD;

[0104] The SVM algorithm module is used to diagnose the faults of the mechanical structure of the disconnector.

[0105] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0106] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0107] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0108] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0109] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A disconnector fault diagnosis method based on Levy-FOX-VMD-SVM, characterized in that: include: The vibration signals of normal operation, inadequate opening and closing, three-phase out of phase and stuck state are extracted from the historical data of the disconnector. The four groups of vibration signals are calculated by VMD algorithm, and the vibration signals are decomposed into k layers of IMF signals. The minimum power spectrum entropy of each layer of IMF signals of the four groups of signals is calculated as the fitness function, and the Levy-FOX algorithm is used to optimize the parameters of VMD. The first group of optimal parameter values ​​obtained are used as the parameters of VMD. The vibration signals are regrouped for VMD, and the characteristic entropy values ​​of each layer of IMF signals are calculated to form a multivariate feature fusion matrix. The PCA algorithm is used to reduce the dimension of the multivariate feature fusion matrix and select the principal components to obtain a three-dimensional reconstruction space. The fault detection module is built using the SVM classifier, and the vibration signal of the mechanical structure of the disconnector to be tested is input. The fault detection module performs eigenvalue matching and outputs the fault detection result.

2. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 1 is characterized in that: The method of calculating the four groups of vibration signals using the VMD algorithm and decomposing the vibration signals into k layers of IMF signals includes: The IMF signal decomposed by VMD is: u k (t)=A k (t)cos[φ k (t)] Where: A k (t) is the IMF signal amplitude, φ k (t) is the instantaneous phase; The VMD constraint requires that the sum of the bandwidths of each IMF center frequency be minimized. {u k } represents the IMF signal set of each layer; {ω k } represents the center frequency set of each layer of IMF; * is the convolution operation; is the gradient operation; δ(t) is the Dirac function; j represents the imaginary unit; f(t) represents the IMF signal of each layer as a function of time, α is the penalty factor; L is the Lagrangian operator, λ is the Lagrangian multiplier; λ(t) is the function of the Lagrangian multiplier with respect to time.

3. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 2 is characterized in that: The minimum power spectrum entropy of each layer of IMF signal is used as the fitness function to adaptively determine the optimal value of the IMF layer number k and the penalty factor α.

4. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 3 is characterized in that: The minimum power spectrum entropy is calculated as follows: S N (f)=|X F (f)| 2 / N Where: f is the original signal, the image signal X represents the IMF components of each layer after VMD decomposition, |X F (f) | represents the Fourier transform of the image signal X, H X is the power spectrum entropy of the graph, ξ i is the percentage of the i-th spectrum power spectrum in the entire spectrum power spectrum, i=1,2...N, where N is the number of sampling points of the IMF signal of each layer.

5. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 4 is characterized in that: The Levy-FOX algorithm is used to optimize the parameters of VMD, including: Step 4.1, set the number of fox groups N = 50, the initial fox search step length is fixed to d = 100, the maximum number of iterations is set to R = 50, the random number r∈[0,1], and the Levy flight function is used to optimize the FOX algorithm. The Levy flight function processes the search step length β of the FOX algorithm as follows to make it conform to the Levy distribution; u and c are normal random variables; β＝1.5； Step 4.2, set the number of VMD decomposition levels k to range from 3 to 10, and optimize the penalty factor α from 1 to 5000; Step 4.3, obtain a random number r, and the fox population starts to perform jumping foraging or random walking foraging. If r ≥ 0.5, jump foraging is performed; if r ≤ 0.5, random walking foraging is performed; Step 4.4, when the fox finds a position with a smaller power spectrum entropy, update the optimal value. If the maximum number of iterations has not been reached, repeat step 4.

3. After the maximum number of iterations is reached, output the optimal values ​​of α and k as the first set of optimal parameter values. The fox position update formula is as follows: Where: Jump is the jumping height of each fox in the algorithm, c and γ are random numbers between [0,1], and X it+1 For the latest location of the fox, X (it) The position of the fox before the update.

6. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 5 is characterized in that: The characteristic entropy values ​​include approximate entropy, sample entropy, fuzzy entropy, Kolmogorov entropy, permutation entropy, modified conditional entropy, distribution entropy, spectral entropy, dispersion entropy, symbolic dynamic entropy, phase entropy, bubble entropy, grid distribution entropy, attention entropy and range entropy.

7. The disconnector fault diagnosis method based on Levy-FOX-VMD-SVM according to claim 6 is characterized in that: The fault detection part includes building a fault detection module using an SVM classifier, inputting the characteristic values ​​of four faults in the PCA three-dimensional reconstruction space into the fault detection module; inputting the vibration signal of the mechanical structure of the isolation switch to be tested, and the fault detection module performs characteristic value matching and outputs the fault detection result.

8. A disconnector fault diagnosis system based on Levy-FOX-VMD-SVM, the system being a system used in a disconnector fault diagnosis method based on Levy-FOX-VMD-SVM as claimed in any one of claims 1 to 7, characterized in that: include: A data acquisition module is used to obtain historical vibration signal data of the disconnector; VMD algorithm module, used to decompose vibration signals into IMF signals; Levy-FOX algorithm module, used to optimize parameters of VMD; The SVM algorithm module is used to diagnose the faults of the mechanical structure of the disconnector.

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images