A method and system for identifying submarine cable vibration based on grey wolf optimization VMD

Abstract

The application discloses a submarine cable vibration identification method and system based on grey wolf optimization VMD, comprising the following steps: acquiring a submarine cable vibration signal, and obtaining an optimal parameter group of adaptive VMD according to a grey wolf optimization algorithm; performing VMD decomposition on the submarine cable vibration signal according to the optimal parameter group to obtain an IMF component containing multi-order modal parameter information; performing singular value decomposition on each order IMF component, and performing submarine cable state identification according to obtained singular value characteristic vectors of each order. The application can improve the precision of submarine cable vibration signal identification, and improve the reliability and accuracy of submarine cable vibration signal identification.

Description

Technical Field

[0001] This invention relates to the field of submarine cable condition identification, specifically to a submarine cable vibration identification method and system based on Grey Wolf Optimized VMD. Background Technology

[0002] To ensure the safety of submarine cables and maximize their carrying capacity, cable condition identification technology is crucial for the transmission of new energy in marine power transmission systems. However, due to the long distances the cables are laid and the lack of effective preventative measures, incidents such as fishing boats and other vessels arbitrarily anchoring or mooring in concentrated cable corridors often cause damage to the cables. Furthermore, excessively high insulation temperatures can lead to insulation material fatigue and shortened lifespan. Therefore, cable condition identification plays a vital role in ensuring the safe transportation of submarine energy. However, existing cable condition identification algorithms lack accuracy and are difficult to use for quantitative analysis. Therefore, a more accurate cable condition identification method is urgently needed. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of the prior art by proposing a submarine cable vibration identification method and system based on Grey Wolf Optimized VMD, so as to improve the accuracy of submarine cable condition identification.

[0004] In a first aspect, the present invention provides a submarine cable vibration identification method based on gray wolf optimized VMD, comprising:

[0005] The vibration signal of the submarine cable was acquired, and the optimal parameter set for adaptive VMD was obtained based on the Grey Wolf optimization algorithm.

[0006] Based on the optimal parameter set, the submarine cable vibration signal is decomposed using VMD to obtain IMF components containing multi-mode parameter information;

[0007] Singular value decomposition is performed on each order of IMF components, and the submarine cable status is identified based on the obtained singular value feature vectors.

[0008] This invention employs the Grey Wolf optimization algorithm to provide an adaptive optimal parameter set for VMD (Vibration Mode Decomposition). The Grey Wolf optimization algorithm significantly improves the decomposition efficiency and accuracy of VMD, enabling adaptive decomposition. Furthermore, the decomposed intrinsic mode components (IMFs) of each order can better characterize modal information, thereby improving the identification of submarine cable vibration signals, and consequently, enhancing the accuracy, reliability, and precision of submarine cable vibration signal identification. Secondly, singular value decomposition (SVD) is performed on each IMF component. Using SVD for feature extraction not only effectively processes non-stationary signals but also extracts eigenvalues ​​that characterize the structural state. Based on the obtained singular value feature vectors of each order, submarine cable state identification is performed, improving the accuracy, reliability, and precision of submarine cable vibration signal identification.

[0009] Furthermore, the singular value decomposition of each order of IMF components includes:

[0010] Each order of IMF component of the submarine cable vibration signal is decomposed into the product of a left odd matrix, a singular value diagonal matrix, and the transpose of a right odd matrix.

[0011] Based on the singular value diagonal matrices corresponding to each order of IMF components, a singular value eigenvector is obtained, which is composed of non-zero singular value features in each singular value diagonal matrix.

[0012] This invention employs diagonal matrices with corresponding singular values ​​for each order of IMF components to obtain singular value feature vectors that can effectively process non-stationary signals and extract structural states, thereby improving the accuracy of submarine cable vibration signal identification and further enhancing its reliability and precision.

[0013] Furthermore, the step of performing VMD decomposition on the submarine cable vibration signal based on the optimal parameter set to obtain IMF components containing multi-order modal parameter information includes:

[0014] Based on the optimal parameters, the submarine cable vibration signal is decomposed into sub-signals with different center frequencies of multiple orders;

[0015] The frequencies corresponding to each order of sub-signals are obtained based on the Hilbert spectrum and the frequencies of each order of IMF components in the hybrid operator set.

[0016] Calculate the bandwidth of each IMF component based on the frequency and the gradient norm of the demodulated signal;

[0017] Based on each bandwidth, a constrained variational model is established, and the constrained variational model is solved to obtain the IMF components that reflect the modal parameters of each order.

[0018] This invention employs the optimal parameters to decompose the submarine cable vibration signal into multiple sub-signals with different center frequencies. Based on the sub-signals of each order, the Hilbert spectrum, and the frequencies of each order IMF component in the hybrid operator set, the frequencies corresponding to each order IMF component can be better concentrated together, facilitating more thorough mode decomposition. This improves the accuracy of submarine cable vibration signal identification, further enhancing its reliability and precision, and reducing the error in submarine cable vibration signal identification.

[0019] Furthermore, the establishment of a constrained variational model based on each bandwidth is specifically as follows:

[0020] A constrained variational model is established with the constraint that the sum of all IMF components equals the original submarine cable vibration signal and the objective function being to minimize the sum of the bandwidths of all IMF components.

[0021] Furthermore, the process of solving the constrained variational model to obtain the IMF components reflecting the modal parameters of each order is specifically as follows:

[0022] Based on the penalty factor, the constrained variational model is continuously updated and optimized using the Lagrange multiplier method and the alternating direction multiplier method until a feasible solution that satisfies the constraints is obtained as the final IMF component; wherein, the optimal parameter set includes the penalty factor.

[0023] This invention employs the Lagrange multiplier method and the alternating direction multiplier method to continuously update and optimize the constrained variational model, which can divide the submarine cable vibration signal into multi-order IMF components. The original submarine cable vibration signal is distinguished according to different natural frequencies, so that each order IMF component has its own natural frequency, which facilitates the identification of submarine cable vibration signals, thereby improving the accuracy of submarine cable vibration signal identification, further improving the reliability and accuracy of submarine cable vibration signal identification, and reducing the error of submarine cable vibration signal identification.

[0024] Furthermore, the optimal parameter set for adaptive VMD obtained according to the Grey Wolf optimization algorithm includes:

[0025] The parameters of the gray wolf optimization algorithm are initialized, and the first fitness function of the VMD is obtained according to the center frequency method. Based on the first fitness function, the optimal first candidate position information of multiple gray wolves in the wolf pack is obtained.

[0026] Based on the dimensional learning-based hunting search strategy, the second candidate position information of each gray wolf is obtained, and the fitness values ​​of each first candidate position information and each second candidate position information are compared to update the position of each gray wolf. After satisfying the number of iterations, the optimal parameter set of adaptive VMD is obtained.

[0027] Furthermore, the first fitness function of the VMD is obtained according to the center frequency method, and the optimal first candidate position information of multiple gray wolves in the wolf pack is obtained according to the first fitness function, including:

[0028] Based on the center frequency method, the first fitness function of VMD is obtained. Based on the first fitness function, the optimal initial positions of multiple alpha wolves in the wolf pack are evaluated. Based on the optimal initial positions, the optimal first candidate position information of gray wolves is calculated in sequence to obtain the optimal first candidate position information of each gray wolf.

[0029] Furthermore, the fitness values ​​of each first candidate location information and each second candidate location information are compared to update the location of each gray wolf, specifically as follows:

[0030] Based on the second fitness function, calculate the first fitness value of the first candidate location information and the second fitness value of the second candidate location information respectively;

[0031] When the first fitness value is less than the second fitness value, the position of the corresponding gray wolf is updated according to the first candidate position information corresponding to the first fitness value;

[0032] Otherwise, the position of the corresponding gray wolf is updated based on the second candidate position information corresponding to the second fitness value.

[0033] Preferably, the parameters for initializing the gray wolf optimization algorithm include: the number of wolf pack members, the number of prey, the range for searching prey, and the initial position of the wolf pack.

[0034] Secondly, the present invention provides a submarine cable vibration identification system based on Grey Wolf Optimized VMD, comprising:

[0035] The optimal parameter acquisition module is used to acquire submarine cable vibration signals and obtain the optimal parameter set for adaptive VMD based on the Grey Wolf optimization algorithm.

[0036] The IMF component calculation module is used to perform VMD decomposition on the submarine cable vibration signal according to the optimal parameter set to obtain IMF components containing multi-order modal parameter information.

[0037] The feature vector calculation module is used to perform singular value decomposition on IMF components of each order, and to identify the status of submarine cables based on the obtained singular value feature vectors of each order. Attached Figure Description

[0038] Figure 1 This is a schematic diagram of the process for identifying submarine cable vibration based on Grey Wolf Optimized VMD provided in an embodiment of the present invention;

[0039] Figure 2This is a schematic diagram of the process of calculating the optimal parameter set for adaptive VMD using the IGWO algorithm provided in this embodiment of the invention;

[0040] Figure 3 This is a schematic diagram illustrating the relationship between the number of iterations and the fitness value of the optimal parameter set for calculating submarine cable vibration signals based on the IGWO algorithm, provided in an embodiment of the present invention.

[0041] Figure 4 This is a schematic diagram of the IMF components of each order after decomposing the submarine cable vibration signal based on IGWO-VMD, provided in an embodiment of the present invention.

[0042] Figure 5 This is a schematic diagram of the structure of the submarine cable vibration identification system based on Grey Wolf Optimized VMD provided in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] See Figure 1 This is a flowchart illustrating the submarine cable vibration identification based on Grey Wolf Optimized VMD provided in an embodiment of the present invention, including steps S11 to S13, specifically:

[0045] Step S11: Obtain the submarine cable vibration signal and obtain the optimal parameter set for adaptive VMD based on the Grey Wolf optimization algorithm.

[0046] Preferably, the grey wolf optimization algorithm is the improved grey wolf optimizer (IGWO) algorithm.

[0047] The optimal parameter set for adaptive VMD (Variational Mode Decomposition) is obtained according to the Grey Wolf Optimizer (GWO) algorithm, including: initializing the parameters of the Grey Wolf Optimizer algorithm, obtaining the first fitness function of the VMD according to the center frequency method, and obtaining the optimal first candidate position information of multiple grey wolves in the wolf pack according to the first fitness function.

[0048] Specifically, based on the IGWO algorithm, the optimal decomposition level and penalty factor of adaptive VMD can be obtained each time. VMD decomposes the submarine cable vibration signal into multi-order IMF components according to the optimal decomposition level and penalty factor. Each order IMF component contains information on different modal frequencies, so that the modal characteristics of each order of the signal can be revealed. Finally, singular value decomposition is performed on each order IMF component. By obtaining the singular values ​​of each order, feature values ​​are extracted. Based on the different features extracted, it is possible to distinguish whether the submarine cable is in a normal state or in a damaged state such as anchor impact, so that the model can more stably and accurately identify the state of the submarine cable.

[0049] Preferably, the parameters for initializing the gray wolf optimization algorithm include: the number of wolf pack members, the number of prey, the range for searching prey, and the initial position of the wolf pack.

[0050] Preferably, the initial position of the gray wolf can be represented as:

[0051] X i,j =l i +rand j [0,1]×(u j -l i ), i∈[1,N],j∈[1,D], (1)

[0052] Where D represents the number of prey, specifically the optimal decomposition level and penalty factor for unknown parameters in VMD, [l i ,u j [] represents the range of prey, N represents the number of wolf pack members, and rand j [0,1] is a random function that generates random numbers on [0,1].

[0053] Preferably, the number of prey is 2.

[0054] Preferably, the wolf pack has 20 members.

[0055] See Figure 2 This is a flowchart illustrating the process of calculating the optimal parameter set for adaptive VMD using the IGWO algorithm provided in this embodiment of the invention. The IGWO algorithm includes steps S101 to S109, specifically:

[0056] Step S101: Initialize the parameters of the Grey Wolf optimization algorithm.

[0057] Step S102: Determine if the number of iterations has been exceeded. If the number of iterations has not been exceeded, proceed to step S103; otherwise, end the process.

[0058] Step S103: Compare the fitness values ​​of all gray wolves in the wolf pack, and designate the top 3 gray wolves with the best fitness values ​​as the leader wolves. The position information of the 3 leader wolves corresponds to the optimal initial position.

[0059] Step S104: Determine whether the number of gray wolves at the updated position is less than the maximum wolf pack size. If the number of gray wolves at the updated position is less than the maximum wolf pack size, proceed to step S105; otherwise, proceed to step S105.

[0060] Step S105: Increment the iteration count by 1, and proceed to step S102.

[0061] Step S106: Calculate the first candidate position information of the gray wolf.

[0062] Step S107: Calculate the second candidate position information of the gray wolf.

[0063] Step S108: By comparing the fitness values ​​of the first candidate position information and the second candidate position information, the position of the gray wolf is updated, and the corresponding parameters are selected as the final optimal parameter set.

[0064] Step S109: Increment the number of gray wolves whose positions have been updated by 1, and proceed to step S104.

[0065] Specifically, the first fitness function of the VMD is obtained according to the center frequency method, and the optimal first candidate position information of multiple gray wolves in the wolf pack is obtained according to the first fitness function, including: obtaining the first fitness function of the VMD according to the center frequency method, evaluating the optimal initial position of multiple alpha wolves in the wolf pack according to the first fitness function, and calculating the optimal first candidate position information of gray wolves in turn according to the optimal initial position, thereby obtaining the optimal first candidate position information of each gray wolf.

[0066] Preferably, the first candidate location information can be represented as:

[0067]

[0068]

[0069] Where t is the iteration number, X α (), X β () and X δ ( ) represent the optimal initial positions of the three alpha wolves in the wolf pack. i1 A i2 and A i3 These are the coefficients of the three alpha wolves, D(t) is also a coefficient parameter, and X... i1 (), X i2 () and X i3() represent the location information updated based on the positions of the three alpha wolves. i-GWO (+1) represents the best first candidate position information for the gray wolf in the next iteration.

[0070] It's worth noting that the first candidate location information for each gray wolf is obtained through the leadership of the three alpha wolves. In other words, each gray wolf's first candidate location information for the next iteration is obtained with the help of the three alpha wolves. In the Dimension Learning-based Hunting (DLH) strategy, each gray wolf learns from different neighbors and randomly selected individuals from the overall wolf pack matrix (N rows, D columns) to update its own location information and generate second candidate location information.

[0071] Specifically, based on the dimensional learning-based hunting search strategy, the second candidate position information of each gray wolf is obtained, and the fitness values ​​of each first candidate position information and each second candidate position information are compared to update the position of each gray wolf. After satisfying the number of iterations, the optimal parameter set of adaptive VMD is obtained.

[0072] In the IGWO algorithm, the calculation process for the second candidate position information corresponding to the gray wolf is as follows: calculate the distance radius between the first candidate position information and the remaining gray wolves to obtain the candidate wolf set. The distance radius and the candidate wolf set can be represented as follows:

[0073] R i (t)=||X i (t)-X i-GWO (t+1)||, (4)

[0074] N i (t)={X j (t)|D i (X i (t),X j (t))≤R i (t),X j (t)∈Pop}, (5)

[0075] Where Pop is an N x D matrix representing the overall wolf pack, and D... i (X i (t),X j (t) represents the distance between a gray wolf and other gray wolves in the wolf pack at time t.

[0076] Preferably, the second candidate location information of the gray wolf can be represented as:

[0077] X i-DLH,d (t+1)=X i,d (t)+rand×(X n,d(t)-X r,d (t)), (6)

[0078] Among them, X n,d (t) is from N i X is a randomly selected gray wolf neighbor in (t). r,d (t) is randomly selected from Pop.

[0079] Specifically, the fitness values ​​of each first candidate position information and each second candidate position information are compared to update the position of each gray wolf. Specifically, according to the second fitness function, the first fitness value of the first candidate position information and the second fitness value of the second candidate position information are calculated respectively. When the first fitness value is less than the second fitness value, the position of the corresponding gray wolf is updated according to the first candidate position information corresponding to the first fitness value; otherwise, the position of the corresponding gray wolf is updated according to the second candidate position information corresponding to the second fitness value.

[0080] Preferably, the position information of the gray wolf in the next iteration can be updated as follows:

[0081]

[0082] Where f(·) is the fitness function. See also Figure 3 This is a schematic diagram illustrating the relationship between the number of iterations and the fitness value of the optimal parameter set for calculating submarine cable vibration signals based on the IGWO algorithm, provided in an embodiment of the present invention. As the number of iterations increases, the fitness value decreases, and after the 5th iteration, the fitness value converges.

[0083] Preferably, the number of iterations is set to 5.

[0084] It is worth noting that after each iteration, a set of optimal position information of the wolf pack can be obtained as the optimal parameter set. The optimal parameter set corresponds to the optimal number of decomposition layers and penalty factor required by VMD. After the number of iterations is completed, the final optimal parameter set is obtained, and the coefficient parameters are updated in each iteration.

[0085] Step S12: Based on the optimal parameter set, perform VMD decomposition on the submarine cable vibration signal to obtain IMF components containing multi-order modal parameter information.

[0086] The step of performing VMD decomposition on the submarine cable vibration signal based on the optimal parameter set to obtain IMF components containing multi-order modal parameter information includes steps S121 to S124, specifically:

[0087] Step S121: Based on the optimal parameters, decompose the submarine cable vibration signal into sub-signals with different center frequencies of multiple orders.

[0088] Step S122: Based on the sub-signals of each order and the Hilbert spectrum, and the frequencies of the IMF components of each order in the mixing operator set, obtain the frequencies corresponding to the sub-signals of each order.

[0089] Preferably, the frequency corresponding to a sub-signal is:

[0090]

[0091] in, For the hybrid operator, u m (t) is a sub-signal, and δ(t) is the Dirac distribution function.

[0092] Step S123: Calculate the bandwidth of each order IMF component based on the gradient norm of each frequency and the demodulated signal.

[0093] Preferably, the bandwidth corresponding to one IMF component can be expressed as:

[0094]

[0095] in, This indicates the partial derivative.

[0096] Step S124: Based on each bandwidth, establish a constrained variational model and solve the constrained variational model to obtain the IMF components that reflect the modal parameters of each order.

[0097] Specifically, a constrained variational model is established based on each bandwidth. The constraint condition is that the sum of the IMF components of each order is equal to the original submarine cable vibration signal, and the objective function is to minimize the sum of the bandwidths of the IMF components of each order.

[0098] Preferably, the constrained variational model can be expressed as:

[0099]

[0100] Where S is the original submarine cable vibration signal, and k is the optimal number of decomposition layers.

[0101] See Figure 4 This is a schematic diagram of the IMF components of the submarine cable vibration signal after decomposition based on IGWO-VMD according to an embodiment of the present invention. It includes an original submarine cable vibration signal and 5 IMF components. The number of decomposition layers of the IMF components is obtained by adaptive optimization calculation using the IGWO algorithm.

[0102] This invention employs the optimal parameters to decompose the submarine cable vibration signal into multiple sub-signals with different center frequencies. Based on the sub-signals of each order, the Hilbert spectrum, and the frequencies of each order IMF component in the hybrid operator set, the frequencies corresponding to each order IMF component can be better concentrated together, facilitating more thorough mode decomposition. This improves the accuracy of submarine cable vibration signal identification, further enhancing its reliability and precision, and reducing the error in submarine cable vibration signal identification.

[0103] Solving the constrained variational model yields the IMF components that reflect the modal parameters of each order. Specifically, based on the penalty factor, the constrained variational model is continuously updated and optimized using the Lagrange multiplier method and the alternating direction multiplier method until a feasible solution that satisfies the constraints is obtained as the final IMF component; wherein, the optimal parameter set includes the penalty factor.

[0104] Preferably, according to the Lagrange multiplier method, the constrained variational model can be transformed into:

[0105]

[0106] Where α and λ(t) are the penalty factor and Lagrange multiplier, respectively.

[0107] Preferably, according to the alternating direction multiplier method, the sub-signal can be transformed into:

[0108]

[0109] Among them, f f(t) For the Fourier transform of the original submarine cable vibration signal, f λ(t) This is the Fourier transform of the Lagrange multiplier operator, where ω is the frequency value. n denoted as the center frequency of each order of IMF.

[0110] It is worth noting that after repeating the process until the constraints are met, the final decomposition result of VMD can be obtained. VMD divides the submarine cable vibration signal into multi-order IMF components, each of which has its own natural frequency, thus distinguishing the original vibration signal according to different natural frequencies.

[0111] This invention employs the Lagrange multiplier method and the alternating direction multiplier method to continuously update and optimize the constrained variational model, which can divide the submarine cable vibration signal into multi-order IMF components. The original submarine cable vibration signal is distinguished according to different natural frequencies, so that each order IMF component has its own natural frequency, which facilitates the identification of submarine cable vibration signals, thereby improving the accuracy of submarine cable vibration signal identification, further improving the reliability and accuracy of submarine cable vibration signal identification, and reducing the error of submarine cable vibration signal identification.

[0112] Step S13: Perform singular value decomposition on each order of IMF components, and identify the submarine cable status based on the obtained singular value feature vectors of each order.

[0113] The process of performing Singular Value Decomposition (SVD) on each order of IMF components includes: decomposing each order of IMF components of the submarine cable vibration signal into the product of a left odd matrix, a singular value diagonal matrix, and the transpose of a right odd matrix; and obtaining a singular value eigenvector composed of non-zero singular value features in each singular value diagonal matrix based on the obtained singular value diagonal matrices corresponding to each order of IMF components.

[0114] Preferably, the singular value decomposition of an IMF component can be expressed as:

[0115] X=UΣV T (13)

[0116] Where X is an M×1 IMF component, U and V are left odd matrices and right odd matrices, respectively, ∑ is a singular value diagonal matrix, ∑=diag[σ,0,...,0], and σ is the singular value of this order of IMF component.

[0117] See Table 1 for the results of singular value feature extraction using the IGWO-VMD-SVD method provided in this embodiment of the invention.

[0118] Table 1. Results of Singular Value Feature Extraction Using the IGWO-VMD-SVD Method

[0119]

[0120] In the table, the number of decomposition layers of the submarine cable vibration signal obtained by adaptively using the IGWO algorithm is 5. It can be seen that there is a significant difference in the singular values ​​of the submarine cable under normal conditions and under anchor damage conditions. The submarine cable condition can be identified based on the singular value feature vector obtained from the final decomposition.

[0121] This invention employs diagonal matrices with corresponding singular values ​​for each order of IMF components to obtain singular value feature vectors that can effectively process non-stationary signals and extract structural states, thereby improving the accuracy of submarine cable vibration signal identification and further enhancing its reliability and precision.

[0122] See Figure 5 This is a schematic diagram of the structure of the submarine cable vibration identification system based on Grey Wolf Optimized VMD provided in an embodiment of the present invention, including: an optimal parameter acquisition module 51, an IMF component calculation module 52, and a feature vector calculation module 53.

[0123] The optimal parameter acquisition module 51 is used to acquire the submarine cable vibration signal and obtain the optimal parameter set for adaptive VMD according to the Grey Wolf optimization algorithm.

[0124] The optimal parameter set for adaptive VMD is obtained according to the gray wolf optimization algorithm, including: initializing the parameters of the gray wolf optimization algorithm, obtaining the first fitness function of the VMD according to the center frequency method, and obtaining the optimal first candidate position information of multiple gray wolves in the wolf pack according to the first fitness function.

[0125] Preferably, the parameters for initializing the gray wolf optimization algorithm include: the number of wolf pack members, the number of prey, the range for searching prey, and the initial position of the wolf pack.

[0126] Specifically, the first fitness function of the VMD is obtained according to the center frequency method, and the optimal first candidate position information of multiple gray wolves in the wolf pack is obtained according to the first fitness function, including: obtaining the first fitness function of the VMD according to the center frequency method, evaluating the optimal initial position of multiple alpha wolves in the wolf pack according to the first fitness function, and calculating the optimal first candidate position information of gray wolves in turn according to the optimal initial position, thereby obtaining the optimal first candidate position information of each gray wolf.

[0127] Furthermore, based on the dimensional learning-based hunting search strategy, the second candidate position information of each gray wolf is obtained, and the fitness values ​​of each first candidate position information and each second candidate position information are compared to update the position of each gray wolf. After satisfying the number of iterations, the optimal parameter set of adaptive VMD is obtained.

[0128] Specifically, the fitness values ​​of each first candidate position and each second candidate position are compared to update the position of each gray wolf, as follows:

[0129] Based on the second fitness function, calculate the first fitness value of the first candidate location information and the second fitness value of the second candidate location information respectively;

[0130] When the first fitness value is less than the second fitness value, the position of the corresponding gray wolf is updated according to the first candidate position information corresponding to the first fitness value;

[0131] Otherwise, the position of the corresponding gray wolf is updated based on the second candidate position information corresponding to the second fitness value.

[0132] The IMF component calculation module 52 is used to perform VMD decomposition on the submarine cable vibration signal according to the optimal parameter set to obtain IMF components containing multi-order modal parameter information.

[0133] The step of performing VMD decomposition on the submarine cable vibration signal based on the optimal parameter set to obtain IMF components containing multi-order modal parameter information includes: a sub-signal calculation submodule 121, a frequency calculation submodule 122, a bandwidth calculation submodule 123, and an IMF component calculation submodule 124, specifically:

[0134] The sub-signal calculation submodule 121 is used to decompose the submarine cable vibration signal into multiple sub-signals with different center frequencies according to the optimal parameters.

[0135] The frequency calculation submodule 122 is used to obtain the frequency corresponding to each order sub-signal based on the frequency of each order sub-signal, the Hilbert spectrum, and the frequency of each order IMF component in the hybrid operator set.

[0136] The bandwidth calculation submodule 123 is used to calculate the bandwidth of each order of IMF components based on each frequency and the gradient norm of the demodulated signal.

[0137] The IMF component calculation submodule 124 is used to establish a constrained variational model based on each bandwidth, and solve the constrained variational model to obtain the IMF components that reflect the modal parameters of each order.

[0138] Specifically, a constrained variational model is established based on each bandwidth. The constraint condition is that the sum of the IMF components of each order is equal to the original submarine cable vibration signal, and the objective function is to minimize the sum of the bandwidths of the IMF components of each order.

[0139] Solving the constrained variational model yields the IMF components that reflect the modal parameters of each order. Specifically, based on the penalty factor, the constrained variational model is continuously updated and optimized using the Lagrange multiplier method and the alternating direction multiplier method until a feasible solution that satisfies the constraints is obtained as the final IMF component; wherein, the optimal parameter set includes the penalty factor.

[0140] The feature vector calculation module 53 is used to perform singular value decomposition on IMF components of each order, and to identify the state of submarine cables based on the obtained singular value feature vectors of each order.

[0141] The singular value decomposition of each order of IMF components includes: decomposing each order of IMF components of the submarine cable vibration signal into the product of a left odd matrix, a singular value diagonal matrix, and the transpose of a right odd matrix; and obtaining a singular value feature vector composed of non-zero singular value features in each singular value diagonal matrix based on the obtained singular value diagonal matrices corresponding to each order of IMF components.

[0142] This invention employs the Grey Wolf Optimization (IGWO) algorithm to provide an adaptive optimal parameter set for VMD. The IGWO algorithm significantly improves the decomposition efficiency and accuracy of VMD, enabling adaptive decomposition. This not only solves problems such as difficulty in parameter determination and low decomposition accuracy but also makes optimization more stable and accurate. For nonlinear and non-stationary signals, VMD decomposition exhibits better mode separation performance compared to other methods, showing better extraction of modes for submarine cable structures. This improves the identification of submarine cable vibration signals, thereby enhancing the accuracy, reliability, and precision of submarine cable vibration signal identification. Secondly, singular value decomposition (SVD) is performed on each order of IMF components. SVD is then used for feature extraction, selecting singular values ​​that stably represent key information as feature values. Thus, the singular values ​​of each order of IMF components can represent important modal information of the corresponding order, demonstrating strong discriminative ability and improving the accuracy, reliability, and precision of submarine cable vibration signal identification.

[0143] Those skilled in the art will understand that embodiments of this application may also include computer program products. Therefore, this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0145] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0146] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0147] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A submarine cable vibration identification method based on grey wolf optimization VMD, characterized in that, The method comprises the steps of: acquiring a submarine cable vibration signal, and obtaining an optimal parameter set of adaptive VMD according to a grey wolf optimization algorithm, wherein the step of obtaining the optimal parameter set of adaptive VMD according to the grey wolf optimization algorithm comprises the steps of: initializing parameters of the grey wolf optimization algorithm, obtaining a first fitness function of the VMD according to a center frequency method, and obtaining first candidate position information of a plurality of grey wolves in a wolf pack according to the first fitness function; obtaining second candidate position information of each grey wolf according to a hunting search strategy based on dimension learning, comparing fitness values of the first candidate position information and the second candidate position information, and updating positions of the grey wolves to obtain the optimal parameter set of adaptive VMD after a number of iterations; performing VMD decomposition on the submarine cable vibration signal according to the optimal parameter set to obtain IMF components containing multi-order modal parameter information; performing singular value decomposition on each order of IMF component, and performing submarine cable state recognition according to obtained singular value feature vectors of each order.

2. The grey wolf optimization VMD-based method for identifying the vibration of the submarine cable according to claim 1, wherein, The singular value decomposition on each order of IMF component comprises the steps of: decomposing each order of IMF component of the submarine cable vibration signal into a product of a left singular matrix, a singular value diagonal matrix, and a transpose of a right singular matrix; obtaining singular value feature vectors composed of non-zero singular values in each singular value diagonal matrix according to each singular value diagonal matrix corresponding to each order of IMF component.

3. The grey wolf optimization VMD-based method for identifying the vibration of the submarine cable according to claim 1, wherein, The VMD decomposition on the submarine cable vibration signal according to the optimal parameter set to obtain IMF components containing multi-order modal parameter information comprises the steps of: decomposing the submarine cable vibration signal into a plurality of sub-signals with different center frequencies according to the optimal parameter; obtaining frequencies corresponding to each order of sub-signal according to each order of sub-signal and Hilbert spectrum and frequencies of each order of IMF component in a mixing operator; calculating bandwidths of each order of IMF component according to each frequency and a gradient norm of a demodulation signal; establishing a constrained variational model according to each bandwidth, and solving the constrained variational model to obtain IMF components reflecting each order of modal parameter.

4. The grey wolf optimization VMD-based method for identifying the vibration of the submarine cable according to claim 3, wherein, The establishment of the constrained variational model according to each bandwidth comprises the steps of: establishing a constrained variational model with a constraint condition that a sum of each order of IMF component is equal to an original submarine cable vibration signal and a target function that minimizes a total sum of bandwidths of each order of IMF component.

5. The grey wolf optimization VMD-based method for identifying the vibration of the submarine cable according to claim 3, wherein, The solving of the constrained variational model to obtain IMF components reflecting each order of modal parameter comprises the steps of: continuously updating and optimizing the constrained variational model by a Lagrange multiplier method and an alternating direction multiplier method according to a penalty factor until a feasible solution satisfying the constraint condition is obtained as the final IMF component; wherein the optimal parameter set comprises the penalty factor.

6. The grey wolf optimization VMD-based method for identifying the vibration of a submarine cable according to claim 1, wherein, The obtaining of the first fitness function of the VMD according to the center frequency method and the obtaining of the first candidate position information of the plurality of grey wolves in the wolf pack according to the first fitness function comprise the steps of: According to the center frequency method, a first fitness function of the VMD is obtained, the optimal initial positions of multiple leader wolves in the wolf pack are evaluated according to the first fitness function, and the optimal first candidate position information of each gray wolf is calculated in turn according to the optimal initial positions, so as to obtain the optimal first candidate position information of each gray wolf.

7. The grey wolf optimization VMD-based method for identifying the vibration of a submarine cable according to claim 1, wherein, The fitness values of the first candidate position information and the second candidate position information are compared, and the position of each gray wolf is updated, specifically as follows: According to the second fitness function, the first fitness value of the first candidate position information and the second fitness value of the second candidate position information are calculated respectively. When the first fitness value is less than the second fitness value, the position of the corresponding gray wolf is updated according to the first candidate position information corresponding to the first fitness value. Otherwise, the position of the corresponding gray wolf is updated according to the second candidate position information corresponding to the second fitness value.

8. The grey wolf optimization VMD-based method for identifying the vibration of a submarine cable according to claim 1, wherein, The parameters of the gray wolf optimization algorithm are initialized, including the number of wolf pack members, the number of prey, the range of searching prey, and the initial position of the wolf pack.

9. A submarine cable vibration identification system based on grey wolf optimization VMD, characterized in that, Including: The optimal parameter acquisition module is configured to acquire a submarine cable vibration signal and obtain an optimal parameter set of adaptive VMD according to a gray wolf optimization algorithm, wherein the optimal parameter set of adaptive VMD is obtained according to the gray wolf optimization algorithm, including initializing parameters of the gray wolf optimization algorithm, obtaining a first fitness function of the VMD according to a center frequency method, and obtaining optimal first candidate position information of multiple gray wolves in a wolf pack according to the first fitness function; obtaining second candidate position information of each gray wolf according to a hunting search strategy based on dimension learning, comparing the fitness values of the first candidate position information and the second candidate position information, updating the position of each gray wolf, and obtaining the optimal parameter set of adaptive VMD after satisfying the number of iterations; The IMF component calculation module is configured to perform VMD decomposition on the submarine cable vibration signal according to the optimal parameter set to obtain IMF components containing multi-order modal parameter information. The feature vector calculation module is configured to perform singular value decomposition on each order IMF component, and perform submarine cable state recognition according to the obtained singular value feature vector of each order.

Images