An insulation oil ultrasonic feature extraction method and device based on WPEE-KPCA

Abstract

The application discloses an insulating oil ultrasonic feature extraction method and device based on WPEE-KPCA, and through collecting insulating oil ultrasonic signal data, wavelet packet decomposition of a preset layer number is performed on the insulating oil ultrasonic signal data to obtain a to-be-calculated wavelet packet, energy entropy spectrum is obtained by using signal data in the to-be-calculated wavelet packet, kernel principal component analysis is performed according to the energy entropy spectrum to obtain a decomposition feature result, the contribution rate of the decomposition feature result is calculated, and the ultrasonic feature of the insulating oil ultrasonic signal data is obtained according to the contribution rate. The embodiment of the application can effectively extract features of acquired insulating oil ultrasonic data, ensure the feature extraction result accuracy, and further improve the accuracy of subsequent application of ultrasonic data for oil state diagnosis.

Description

Technical Field

[0001] This invention relates to the field of ultrasonic data analysis and feature extraction, and in particular to a method and apparatus for ultrasonic feature extraction of insulating oil based on WPEE-KPCA. Background Technology

[0002] Ultrasonic testing is a technique that studies the relationship between the acoustic parameters of the tested medium and the tested object to obtain the physicochemical properties of the medium. Due to its non-destructive and rapid characteristics, this technique has been widely used in the structural testing and performance analysis of solid and liquid materials.

[0003] However, when using ultrasonic technology to detect substances, transformer oil is an extremely complex mixture, and its actual ultrasonic measurement is the result of the superposition of multiple signals. Therefore, the ultrasonic data of transformer oil contains both useful information characterizing the aging state of the transformer and invalid information such as background interference and noise signals. This useless information can affect the results of subsequent diagnostics of the transformer oil condition using ultrasonic data. Therefore, the accuracy of directly using ultrasound for oil condition diagnosis is difficult to meet the needs of practical engineering applications. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method and apparatus for ultrasonic feature extraction of insulating oil based on WPEE-KPCA. This method effectively extracts features from the acquired ultrasonic data of insulating oil, ensuring the accuracy of the feature extraction results and thereby improving the accuracy of subsequent applications of ultrasonic data for oil condition diagnosis.

[0005] A first aspect of this invention provides a method for ultrasonic feature extraction of insulating oil based on WPEE-KPCA, the method comprising:

[0006] Acquire ultrasonic signal data of insulating oil;

[0007] The ultrasonic signal data of the frequency band insulating oil is decomposed into wavelet packets with a preset number of layers to obtain the wavelet packets to be calculated.

[0008] The energy entropy spectrum is obtained from the signal data in the wavelet packet to be calculated. Kernel principal component analysis is performed based on the energy entropy spectrum to obtain the decomposition feature results. The contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the insulating oil ultrasonic signal data are obtained based on the contribution rate.

[0009] In one possible implementation of the first aspect, wavelet packet decomposition is performed on the ultrasonic signal data of insulating oil to obtain multiple wavelet packets, including:

[0010] The ultrasonic signal data of insulating oil is decomposed into wavelet packets of a preset number of layers using a decomposition formula, resulting in multiple wavelet packets. The decomposition formula is as follows:

[0011]

[0012] In the formula, The nodal coefficients of the k-th layer node (j+1, 2n) in the wavelet packet decomposition under a high-pass filter are given. h represents the node coefficients (j+1, 2n+1) of the k-th layer in the wavelet packet decomposition under a low-pass filter. 2l-k g 2l-k For high-pass and low-pass filters, Let be the node coefficients of the l-th layer node (j,n) of the wavelet packet.

[0013] In one possible implementation of the first aspect, the energy entropy spectrum is obtained using signal data in the wavelet packet to be computed, including:

[0014] The signal data of each frequency band in the wavelet packet to be calculated are integrated to obtain the energy entropy of each signal data in each frequency band. The formula for calculating the energy entropy is as follows:

[0015]

[0016] In the formula, E j s represents the energy entropy of each frequency band. j (t) represents the j-th signal data in each frequency band;

[0017] The energy entropy spectrum of each frequency band is obtained using the energy entropy of the signal data in each frequency band. The expression for the energy entropy spectrum of each frequency band is as follows:

[0018] q = [E1, E2, ..., E j ]

[0019] In the formula, q represents the energy entropy spectrum, and E j This represents the energy entropy of each frequency band.

[0020] In one possible implementation of the first aspect, nuclear principal component analysis is performed based on the energy entropy spectrum to obtain the analysis results, including:

[0021] The data to be analyzed is obtained based on the energy entropy spectrum of each frequency band, where the expression for the data to be analyzed is:

[0022] Q = [q1, q2, ..., q n ]

[0023] In the formula, n is the number of frequency bands, q i q represents the energy entropy spectrum of the i-th wavelet packet in the wavelet packet to be calculated. i It is a j-dimensional vector;

[0024] The data to be analyzed is mapped to a high-dimensional feature space to obtain the covariance matrix of the data. Then, eigenvalue decomposition is performed on the covariance matrix to obtain the decomposed feature results, which are as follows:

[0025]

[0026] In the formula, λ = [λ1, λ2,..., λ n [C] is the covariance matrix. F , where v is the eigenvalue matrix and v is the corresponding eigenvector matrix.

[0027] In one possible implementation of the first aspect, the contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the insulating oil ultrasonic signal data are obtained based on the contribution rate, including:

[0028] The number of principal components to be calculated is determined based on the eigenvalue matrix in the decomposition feature results. The formula for calculating the number of principal components to be calculated is as follows:

[0029]

[0030] In the formula, V CP (n) represents the number of pivot components, λ i Let N represent the i-th eigenvalue matrix, and N represent the total number of eigenvalue matrices.

[0031] The contribution rate of the principal components to be calculated in the decomposition feature results is used to obtain the first contribution rate. The formula for calculating the contribution rate is as follows:

[0032]

[0033] In the formula, λ1 represents the first eigenvalue matrix, λ i Let represent the i-th eigenvalue matrix, and n represent the number of principal components to be calculated;

[0034] The contribution rate of the principal elements to be calculated is judged. If the contribution rate exceeds the first preset value, the first number of principal elements to be calculated is selected for contribution rate calculation to obtain the second contribution rate. If the second contribution rate is greater than the second preset value, the eigenvalue matrix of the first number of principal elements is determined to be the ultrasonic feature of the insulating oil ultrasonic signal data.

[0035] A second aspect of this invention provides an ultrasonic feature extraction device for insulating oil based on WPEE-KPCA, the device comprising:

[0036] The acquisition module is used to acquire ultrasonic signal data of insulating oil;

[0037] The decomposition module is used to perform wavelet packet decomposition on the ultrasonic signal data of insulating oil to obtain multiple wavelet packets, and select the wavelet packets that meet the preset conditions as the wavelet packets to be calculated.

[0038] The analysis module is used to obtain the energy entropy spectrum from the signal data in the wavelet packet to be calculated, perform kernel principal component analysis based on the energy entropy spectrum to obtain the decomposition feature results, calculate the contribution rate of the decomposition feature results, and obtain the ultrasonic features of the insulating oil ultrasonic signal data based on the contribution rate.

[0039] In one possible implementation of the second aspect, wavelet packet decomposition is performed on the ultrasonic signal data of insulating oil to obtain multiple wavelet packets, including:

[0040] The ultrasonic signal data of insulating oil is decomposed into wavelet packets of a preset number of layers using a decomposition formula, resulting in multiple wavelet packets. The decomposition formula is as follows:

[0041]

[0042] In the formula, The nodal coefficients of the k-th layer node (j+1, 2n) in the wavelet packet decomposition under a high-pass filter are given. h represents the node coefficients (j+1, 2n+1) of the k-th layer in the wavelet packet decomposition under a low-pass filter. 2l-k g 2l-k For high-pass and low-pass filters, Let be the node coefficients of the l-th layer node (j,n) of the wavelet packet.

[0043] In one possible implementation of the second aspect, the energy entropy spectrum is obtained using signal data in the wavelet packet to be computed, including:

[0044] The signal data of each frequency band in the wavelet packet to be calculated are integrated to obtain the energy entropy of each signal data in each frequency band. The formula for calculating the energy entropy is:

[0045]

[0046] In the formula, E j s represents the energy entropy of each frequency band. j (t) represents the j-th signal data in each frequency band;

[0047] The energy entropy spectrum of each frequency band is obtained using the energy entropy of the signal data in each frequency band. The expression for the energy entropy spectrum of each frequency band is as follows:

[0048] q = [E1, E2, ..., E j ]

[0049] In the formula, q represents the energy entropy spectrum, and E j This represents the energy entropy of each frequency band.

[0050] A third aspect of the present invention provides a computer device, comprising:

[0051] Memory, used to store computer programs;

[0052] A processor for executing computer programs to implement the WPEE-KPCA-based ultrasonic feature extraction method for insulating oil, as described in the first aspect.

[0053] A fourth aspect of the present invention provides a storage medium storing a computer program, which, when executed by a processor, implements the steps of the WPEE-KPCA-based ultrasonic feature extraction method for insulating oil as described in the first aspect.

[0054] This invention, through the acquisition of ultrasonic signal data of insulating oil, performs wavelet packet decomposition on the ultrasonic signal data of insulating oil at a preset number of layers to obtain wavelet packets to be calculated. Energy entropy spectra are obtained using the signal data in the wavelet packets to be calculated. Kernel principal component analysis is performed based on the energy entropy spectra to obtain decomposition feature results. The contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the ultrasonic signal data of insulating oil are obtained based on the contribution rate. This invention can effectively extract features from the acquired ultrasonic data of insulating oil, ensuring the accuracy of the feature extraction results, thereby improving the accuracy of subsequent application of ultrasonic data for oil condition diagnosis. Attached Figure Description

[0055] Figure 1 : A flowchart illustrating the configuration method of an embodiment of the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA provided by the present invention;

[0056] Figure 2 : Original ultrasonic images of different oil states in one embodiment of the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA provided by the present invention;

[0057] Figure 3 : KPCA cumulative principal component contribution rate diagram of the energy spectrum of an embodiment of the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA provided by the present invention;

[0058] Figure 4 : A block diagram of an embodiment of the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA provided by the present invention. Detailed Implementation

[0059] 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.

[0060] Please refer to Figure 1 This is a flowchart illustrating an embodiment of the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA provided by the present invention, including steps S101 to S103, each step of which is as follows:

[0061] S101. Collect ultrasonic signal data of insulating oil.

[0062] In this embodiment, ultrasonic data of the insulating oil sample is collected using an ultrasonic detector, wherein the expression for the ultrasonic data is:

[0063] s(t)=(x1,x2,…,x n )

[0064] In the formula, s(t) represents the ultrasonic data, and x n This represents the nth ultrasound data.

[0065] It is worth noting that ultrasonic signal data, also known as ultrasound signal data, is obtained by using an ultrasonic probe to emit ultrasound, while also receiving the reflected, refracted, and diffracted ultrasound signals, and then converting these sound waves into electrical frequency signals.

[0066] S102. Perform wavelet packet decomposition on the ultrasonic signal data of insulating oil with a preset number of layers to obtain the wavelet packet to be calculated.

[0067] In this embodiment, the ultrasonic signal data of insulating oil is decomposed into wavelet packets of a preset number of layers to obtain the wavelet packets to be calculated. Then, the signals s of each frequency band of the last wavelet packet after decomposition are selected. j The integral of the square of (t) over time is the energy entropy.

[0068] It should be noted that the preset number of layers can be selected according to the actual situation. Wavelet packet decomposition can also be called wavelet packet analysis, subband tree analysis, or optimal subband tree structure analysis. Wavelet packet decomposition refers to the analysis of the detailed parts of the input signal using multiple iterations of wavelet transform.

[0069] In some embodiments, step S102 includes, but is not limited to, the following steps:

[0070] The ultrasonic signal data of insulating oil is decomposed into wavelet packets of a preset number of layers using a decomposition formula, resulting in multiple wavelet packets. The decomposition formula is as follows:

[0071]

[0072] In the formula, The nodal coefficients of the k-th layer node (j+1, 2n) in the wavelet packet decomposition under a high-pass filter are given. h represents the node coefficients (j+1, 2n+1) of the k-th layer in the wavelet packet decomposition under a low-pass filter. 2l-k g 2l-k For high-pass and low-pass filters, Let be the node coefficients of the l-th layer node (j,n) of the wavelet packet.

[0073] Specifically, for ultrasonic data s(t), its wavelet packet decomposition can be expressed as:

[0074]

[0075] In the formula, The nodal coefficients of the k-th layer node (j+1, 2n) in the wavelet packet decomposition under a high-pass filter are given. h represents the node coefficients (j+1, 2n+1) of the k-th layer in the wavelet packet decomposition under a low-pass filter. 2l-k g 2l-k For high-pass and low-pass filters, Let be the node coefficients of the l-th layer node (j,n) of the wavelet packet.

[0076] S103: Obtain the energy entropy spectrum using the signal data in the wavelet packet to be calculated, perform kernel principal component analysis based on the energy entropy spectrum to obtain the decomposition feature results, calculate the contribution rate of the decomposition feature results, and obtain the ultrasonic features of the insulating oil ultrasonic signal data based on the contribution rate.

[0077] Specifically, the signals s of each frequency band of the last wavelet packet after decomposition are selected. j The integral of the square of (t) over time is used as the energy entropy. After obtaining the energy entropy of each signal, the energy entropy spectrum is obtained by using the energy entropy of each signal.

[0078] In some embodiments, step S103 includes, but is not limited to, the following steps:

[0079] The signal data of each frequency band in the wavelet packet to be calculated are integrated to obtain the energy entropy of each signal data in each frequency band. The formula for calculating the energy entropy is:

[0080]

[0081] In the formula, E j s represents the energy entropy of each frequency band. j (t) represents the j-th signal data in each frequency band;

[0082] The energy entropy spectrum of each frequency band is obtained using the energy entropy of the signal data in each frequency band. The expression for the energy entropy spectrum of each frequency band is as follows:

[0083] q = [E1, E2, ..., E j ]

[0084] In the formula, q represents the energy entropy spectrum, and E j This represents the energy entropy of each frequency band.

[0085] Specifically, the signals s of each frequency band of the last wavelet packet after decomposition are selected. j The integral of the square of (t) over time is taken as the energy entropy. The expression for the energy entropy of the j-th frequency band is:

[0086]

[0087] In the formula, E j The energy entropy of each frequency band is represented by s. j (t) represents the j-th signal data in each frequency band;

[0088] The energy entropy spectrum of each frequency band is obtained by using the energy entropy of each frequency band. The expression for the energy entropy spectrum of each frequency band is as follows:

[0089] q = [E1, E2, ..., E j ]

[0090] In the formula, q represents the energy entropy spectrum, and E j This represents the energy entropy of each frequency band.

[0091] In some embodiments, step S103 includes, but is not limited to, the following steps:

[0092] The data to be analyzed is obtained from the energy entropy spectrum of the ultrasonic signal data of insulating oil. The expression for the data to be analyzed is as follows:

[0093] Q = [q1, q2, ..., q n ]

[0094] In the formula, n is the number of frequency bands, q i The energy entropy spectrum, q, represents the ultrasonic signal data of insulating oil. i It is a j-dimensional vector;

[0095] The data to be analyzed is mapped to a high-dimensional feature space to obtain the covariance matrix of the data. Then, eigenvalue decomposition is performed on the covariance matrix to obtain the decomposed feature results, which are as follows:

[0096]

[0097] In the formula, λ = [λ1, λ2,..., λ n [C] is the covariance matrix. F , where v is the eigenvalue matrix and v is the corresponding eigenvector matrix.

[0098] Specifically, the dataset to be processed is first determined, namely the energy entropy spectrum of the ultrasonic signal data of insulating oil, expressed as:

[0099] Q = [q1, q2, ..., q n ]

[0100] In the formula, n is the number of frequency bands, q i The energy entropy spectrum, q, represents the ultrasonic signal data of insulating oil. i It is a j-dimensional vector.

[0101] Then, KPCA was performed on each energy entropy spectrum, utilizing a nonlinear function. The data in the energy entropy spectrum of the ultrasonic signal data of insulating oil are mapped to a high-dimensional feature space F. Therefore, the covariance matrix C of the samples in the feature space F is... F It can be represented as:

[0102]

[0103] In the formula, λ = [λ1, λ2,..., λ n [C] is the covariance matrix. F , where v is the eigenvalue matrix and v is the corresponding eigenvector matrix.

[0104] It is worth noting that frequency band, also known as bandwidth, refers to a range in the frequency domain, defined by a low frequency and a high frequency, or it can refer to a radio band or a range of other spectrum.

[0105] In some embodiments, step S103 includes, but is not limited to, the following steps:

[0106] The number of principal components to be calculated is determined based on the eigenvalue matrix in the decomposition feature results. The formula for calculating the number of principal components to be calculated is as follows:

[0107]

[0108] In the formula, V CP (n) represents the number of pivot components, λ i Let N represent the i-th eigenvalue matrix, and N represent the total number of eigenvalue matrices.

[0109] The contribution rate of the principal components to be calculated in the decomposition feature results is used to obtain the first contribution rate. The formula for calculating the contribution rate is as follows:

[0110]

[0111] In the formula, λ1 represents the first eigenvalue matrix, λ i Let represent the i-th eigenvalue matrix, and n represent the number of principal components to be calculated;

[0112] The contribution rate of the principal elements to be calculated is judged. If the contribution rate exceeds the first preset value, the first number of principal elements to be calculated is selected for contribution rate calculation to obtain the second contribution rate. If the second contribution rate is greater than the second preset value, the eigenvalue matrix of the first number of principal elements is determined to be the ultrasonic feature of the insulating oil ultrasonic signal data.

[0113] Specifically, after performing KPCA processing on each energy entropy spectrum, the eigenvalue matrix after eigenvalue decomposition is obtained. The number of principal components to be calculated is first determined based on the eigenvalue matrix from the eigenvalue decomposition results. The formula for calculating the number of principal components to be calculated is as follows:

[0114]

[0115] In the formula, V CP (n) represents the number of pivot components, λ i Let N represent the i-th eigenvalue matrix, and N represent the total number of eigenvalue matrices.

[0116] Then, the contribution rate of the principal components in the decomposition feature results is calculated to obtain the first contribution rate. The formula for calculating the contribution rate is as follows:

[0117]

[0118] In the formula, λ1 represents the first eigenvalue matrix, λ i Let represent the i-th eigenvalue matrix, and n represent the number of principal components to be calculated;

[0119] The contribution rate of the principal elements to be calculated is judged. If the contribution rate exceeds the first preset value, the first number of principal elements to be calculated is selected for contribution rate calculation to obtain the second contribution rate. If the second contribution rate is greater than the second preset value, the eigenvalue matrix of the first number of principal elements is determined to be the ultrasonic feature of the insulating oil ultrasonic signal data.

[0120] It is worth noting that the principal element refers to the element that plays a dominant role in the elimination process. In linear algebra, the principal element is the first non-zero element in each row of a matrix after elementary transformation into a canonical echelon form matrix, which is always 1.

[0121] As an example of this embodiment, the ultrasonic feature extraction method for insulating oil based on WPEE-KPCA is further illustrated below. Oil samples with different aging times were prepared through accelerated thermal aging tests, and the raw ultrasonic data of the transformer oil were obtained by detecting the oil samples using an ultrasonic testing platform. To verify the effectiveness of the feature extraction, SVM was used for diagnostic verification. The specific steps are as follows:

[0122] First, Raman spectra of insulating oil at different aging times were obtained. According to DL / T 984, the prepared oil samples were experimentally divided into four stages (I: Good, II: Operable, III: Caution, IV: Exit from operation). Then, raw ultrasonic data for different oil states were obtained, such as... Figure 2 As shown. Based on WPEE-KPCA extraction of oil ultrasonic features, this invention recommends a 4-layer wavelet packet decomposition, from... Figure 3 It can be seen that the cumulative contribution rate of the first eight principal components exceeds 99%, and the cumulative contribution rate of the first four principal components exceeds 90%. Therefore, this document recommends the first four principal components as WPEE-KPCA ultrasonic features. Finally, SVM was used to verify the WPEE-KPCA ultrasonic features, and the results are shown in Tables 1 and 2 below. The results show that using this method to extract ultrasonic data of insulating oil significantly improves the accuracy of feature extraction.

[0123] Table 1. Diagnostic results using raw ultrasound data as SVM input.

[0124]

[0125] Table 2. Diagnostic results using oil ultrasonic WPEE-KPCA features as SVM input.

[0126]

[0127] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0128] In some embodiments, such as Figure 4 The diagram shows a block diagram of an ultrasonic feature extraction device 400 for insulating oil based on WPEE-KPCA provided in this application embodiment, including: an acquisition module 401, a decomposition module 402, and an analysis module 403, wherein:

[0129] Acquisition module 401 is used to acquire ultrasonic signal data of insulating oil;

[0130] Decomposition module 402 is used to perform wavelet packet decomposition of the insulating oil ultrasonic signal data at a preset number of layers to obtain the wavelet packet to be calculated;

[0131] Analysis module 403 is used to obtain the energy entropy spectrum from the signal data in the wavelet packet to be calculated, perform kernel principal component analysis based on the energy entropy spectrum to obtain the decomposition feature results, calculate the contribution rate of the decomposition feature results, and obtain the ultrasonic features of the insulating oil ultrasonic signal data based on the contribution rate.

[0132] In some embodiments, wavelet packet decomposition is performed on the ultrasonic signal data of insulating oil to obtain multiple wavelet packets, including:

[0133] The ultrasonic signal data of insulating oil is decomposed into wavelet packets of a preset number of layers using a decomposition formula, resulting in multiple wavelet packets. The decomposition formula is as follows:

[0134]

[0135] In the formula, The nodal coefficients of the k-th layer node (j+1, 2n) in the wavelet packet decomposition under a high-pass filter are given. h represents the node coefficients (j+1, 2n+1) of the k-th layer in the wavelet packet decomposition under a low-pass filter. 2l-k g 2l-k For high-pass and low-pass filters, Let be the node coefficients of the l-th layer node (j,n) of the wavelet packet.

[0136] In some embodiments, obtaining the energy entropy spectrum using signal data in the wavelet packet to be computed includes:

[0137] The signal data of each frequency band in the wavelet packet to be calculated are integrated to obtain the energy entropy of each frequency band. The formula for calculating the energy entropy is as follows:

[0138]

[0139] In the formula, E j s represents the energy entropy of each frequency band. j (t) represents the j-th signal data in each frequency band;

[0140] The energy entropy spectrum of the ultrasonic signal data of insulating oil is obtained using the energy entropy of each frequency band. The expression for the energy entropy spectrum of the ultrasonic signal data of insulating oil is as follows:

[0141] q = [E1, E2, ..., E j ]

[0142] In the formula, q represents the energy entropy spectrum, and E j This represents the energy entropy of each frequency band.

[0143] In some embodiments, nuclear principal component analysis is performed based on the energy entropy spectrum to obtain analytical results, including:

[0144] The data to be analyzed is obtained from the energy entropy spectrum of the ultrasonic signal data of insulating oil. The expression for the data to be analyzed is as follows:

[0145] Q = [q1, q2, ..., q n ]

[0146] In the formula, n is the number of frequency bands, q i The energy entropy spectrum, q, represents the ultrasonic signal data of insulating oil. i It is a j-dimensional vector;

[0147] The data to be analyzed is mapped to a high-dimensional feature space to obtain the covariance matrix of the data. Then, eigenvalue decomposition is performed on the covariance matrix to obtain the decomposed feature results, which are as follows:

[0148]

[0149] In the formula, λ = [λ1, λ2,..., λ n [C] is the covariance matrix. F , where v is the eigenvalue matrix and v is the corresponding eigenvector matrix.

[0150] In some embodiments, the contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the insulating oil ultrasonic signal data are obtained based on the contribution rate, including:

[0151] The number of principal components to be calculated is determined based on the eigenvalue matrix in the decomposition feature results. The formula for calculating the number of principal components to be calculated is as follows:

[0152]

[0153] In the formula, V CP (n) represents the number of pivot components, λ i Let N represent the i-th eigenvalue matrix, and N represent the total number of eigenvalue matrices.

[0154] The contribution rate of the principal components to be calculated in the decomposition feature results is used to obtain the first contribution rate. The formula for calculating the contribution rate is as follows:

[0155]

[0156] In the formula, λ1 represents the first eigenvalue matrix, λ i Let represent the i-th eigenvalue matrix, and n represent the number of principal components to be calculated;

[0157] The contribution rate of the principal elements to be calculated is judged. If the contribution rate exceeds the first preset value, the first number of principal elements to be calculated is selected for contribution rate calculation to obtain the second contribution rate. If the second contribution rate is greater than the second preset value, the eigenvalue matrix of the first number of principal elements is determined to be the ultrasonic feature of the insulating oil ultrasonic signal data.

[0158] In one embodiment of this application, a computer device is provided, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above steps. The implementation principle and technical effects of the computer device provided in this embodiment are similar to those of the above method embodiments, and will not be repeated here.

[0159] In one embodiment of this application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, it performs the above steps; the implementation principle and technical effects of the computer-readable storage medium provided in this embodiment are similar to those of the above method embodiments, and will not be repeated here.

[0160] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0161] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0162] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for extracting ultrasonic features of insulating oil based on WPEE-KPCA, characterized in that, include: Acquire ultrasonic signal data of insulating oil; The ultrasonic signal data of the insulating oil is decomposed into wavelet packets of a preset number of layers using a decomposition formula to obtain the wavelet packets to be calculated. The decomposition formula is as follows: In the formula, For wavelet packet decomposition under high-pass filter, the first... k Layer nodes ( The nodal coefficients of ) For wavelet packet decomposition under low-pass filter, the first... k Layer nodes ( The nodal coefficients of ) , For high-pass and low-pass filters, For the wavelet packet Layer nodes ( The nodal coefficients of ) The signal data of each frequency band in the wavelet packet to be calculated are integrated to obtain the energy entropy of each frequency band, wherein the formula for calculating the energy entropy is: In the formula, This represents the energy entropy of each frequency band. The first frequency band represents the frequency band of each frequency band. Signal data; The energy entropy spectrum of each frequency band is obtained using the energy entropy of each frequency band, wherein the expression for the energy entropy spectrum of the insulating oil ultrasonic signal data is: In the formula, Represents the energy entropy spectrum. This represents the energy entropy of each frequency band. Based on the energy entropy spectrum, principal component analysis is performed to obtain the decomposition feature results. The contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the insulating oil ultrasonic signal data are obtained based on the contribution rate.

2. The method for extracting ultrasonic features of insulating oil based on WPEE-KPCA as described in claim 1, characterized in that, The analysis results obtained by performing nuclear principal component analysis based on the energy entropy spectrum include: The data to be analyzed is obtained based on the energy entropy spectrum of the ultrasonic signal data of the insulating oil, wherein the expression of the data to be analyzed is: In the formula, For the number of frequency bands, This represents the energy entropy spectrum of the ultrasonic signal data of the insulating oil. for j dimensional vector; The data to be analyzed is mapped to a high-dimensional feature space to obtain the covariance matrix of the data to be analyzed. The covariance matrix is ​​then subjected to eigenvalue decomposition to obtain the decomposed feature results, wherein the decomposed feature results are: In the formula, Covariance matrix eigenvalue matrix, Its corresponding eigenvector matrix, It is a nonlinear function.

3. The method for extracting ultrasonic features of insulating oil based on WPEE-KPCA as described in claim 2, characterized in that, The calculation of the contribution rate of the decomposition feature result, and the obtaining of the ultrasonic features of the insulating oil ultrasonic signal data based on the contribution rate, includes: The number of principal components to be calculated is determined based on the eigenvalue matrix in the decomposition feature results, wherein the formula for calculating the number of principal components to be calculated is: In the formula, Indicates the number of pivots. Indicates the first eigenvalue matrix, This represents the total number of eigenvalue matrices; Calculate the contribution rate of the principal components in the decomposition feature results to obtain the first contribution rate, wherein the formula for calculating the contribution rate is: In the formula, This represents the first eigenvalue matrix. Indicates the first eigenvalue matrix, Indicates the number of pivot elements to be calculated; The contribution rate of the principal elements in the number of principal elements to be calculated is judged. If the contribution rate exceeds a first preset value, the first number of principal elements in the number of principal elements to be calculated is selected for contribution rate calculation to obtain a second contribution rate. If the second contribution rate is greater than a second preset value, the eigenvalue matrix of the first number of principal elements is determined to be the ultrasonic feature of the insulating oil ultrasonic signal data.

4. A device for ultrasonic feature extraction of insulating oil based on WPEE-KPCA, characterized in that, include: The acquisition module is used to acquire ultrasonic signal data of insulating oil; The decomposition module is used to perform a preset number of wavelet packet decompositions on the ultrasonic signal data of the insulating oil using a decomposition formula to obtain the wavelet packets to be calculated. The decomposition formula is as follows: In the formula, For wavelet packet decomposition under high-pass filter, the first... k Layer nodes ( The nodal coefficients of ) For wavelet packet decomposition under low-pass filter, the first... k Layer nodes ( The nodal coefficients of ) , For high-pass and low-pass filters, For the wavelet packet Layer nodes ( The nodal coefficients of ) The analysis module is used to integrate the signal data of each frequency band in the wavelet packet to be calculated to obtain the energy entropy of each frequency band, wherein the formula for calculating the energy entropy is: In the formula, This represents the energy entropy of each frequency band. The first frequency band represents the frequency band of each frequency band. Signal data; The energy entropy spectrum of the insulating oil ultrasonic signal data is obtained using the energy entropy of each frequency band, wherein the expression for the energy entropy spectrum of the insulating oil ultrasonic signal data is: In the formula, Represents the energy entropy spectrum. This represents the energy entropy of each frequency band. Based on the energy entropy spectrum, principal component analysis is performed to obtain the decomposition feature results. The contribution rate of the decomposition feature results is calculated, and the ultrasonic features of the insulating oil ultrasonic signal data are obtained based on the contribution rate.

5. Characterized in that, obtaining the energy entropy spectrum using the signal data in the wavelet packet to be calculated includes: A computer device, characterized in that it comprises: Memory, used to store computer programs; A processor, configured to implement the WPEE-KPCA-based ultrasonic feature extraction method for insulating oil as described in any one of claims 1 to 3 when executing the computer program.

6. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the WPEE-KPCA-based ultrasonic feature extraction method for insulating oil as described in any one of claims 1 to 3.

Images