On-load tap-changer vibration accompanying signal feature extraction method and device

Abstract

The invention discloses an on-load tap-changer vibration accompanying signal feature extraction method and device, and relates to the technical field of feature signal extraction, and the method comprises the steps: obtaining a vibration accompanying signal during the operation of an on-load tap-changer; performing de-noising processing on the vibration accompanying signal by adopting CEEMDAN to obtain a de-noised vibration signal; segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain segmented vibration signals; extracting time domain features of the segmented vibration signals; performing time-frequency analysis on the denoised vibration signal by adopting synchronous compression transformation to obtain a time-frequency matrix; extracting texture features representing a signal time-frequency energy distribution mode from the time-frequency matrix based on an image texture analysis method; and constructing a multi-dimensional vibration feature vector according to the time domain feature and the texture feature. According to the invention, the signal-to-noise ratio of the signal can be improved, and the comprehensive extraction of the signal features from the two dimensions of the time domain and the frequency domain is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal feature extraction, in particular to a kind of on-load tap changer vibration accompanying signal feature extraction method and device. BACKGROUND

[0002] In power system, on-load tap changer (OLTC) is the key component of transformer, for adjusting tap position when transformer is running, to realize voltage regulation, to ensure the stable operation of power grid. Vibration accompanying signal as a kind of important information that can reflect the mechanical state and operation process of on-load tap changer, its feature extraction has key significance for early fault diagnosis and state evaluation.

[0003] Traditional vibration signal processing method has many limitations when facing the complex vibration accompanying signal of on-load tap changer. On the one hand, a large amount of noise is often mixed in the signal, which may come from environmental interference, electromagnetic interference of electrical equipment, etc., so that the real characteristics of the signal are difficult to accurately extract. On the other hand, vibration signal has time-varying and nonlinear characteristics, and traditional time domain analysis method is difficult to fully reflect its complex characteristics, and frequency domain analysis is difficult to provide time information. SUMMARY

[0004] In view of the low accuracy and the problem of not being comprehensive in extracting vibration accompanying signal in the prior art, the present application provides an on-load tap changer vibration accompanying signal feature extraction method and device, which can improve the signal-to-noise ratio of the signal and realize comprehensive extraction of signal features from two dimensions of time domain and frequency domain. The specific technical solutions are as follows:

[0005] In the first aspect, the present application provides an on-load tap changer vibration accompanying signal feature extraction method, comprising:

[0006] obtaining vibration accompanying signal when on-load tap changer is running;

[0007] using CEEMDAN to denoise the vibration accompanying signal, to obtain denoised vibration signal;

[0008] segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation, to obtain segmented vibration signal;

[0009] extracting time domain features from the segmented vibration signal;

[0010] using synchronous compression transform to analyze the denoised vibration signal in time-frequency domain, to obtain time-frequency matrix;

[0011] extracting texture features representing signal time-frequency energy distribution pattern from the time-frequency matrix based on image texture analysis method;

[0012] A multi-dimensional vibration feature vector is constructed according to the time domain feature and the texture feature.

[0013] Preferably, the denoising process of the vibration associated signal using CEEMDAN includes:

[0014] Add positive and negative Gaussian white noise to the vibration accompanying signal, perform EMD decomposition, and obtain the intrinsic modal components (IMFs) of each order;

[0015] Calculate the correlation coefficient between each order IMF and the original signal, and filter out the IMF components whose correlation coefficient is greater than the preset threshold;

[0016] The filtered IMF components are reconstructed to obtain the denoised vibration signal.

[0017] Preferably, segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation includes:

[0018] Calculating a short-time energy sequence of the denoised vibration signal through a sliding window;

[0019] The OTSU algorithm is used to select a dynamic segmentation threshold for the short-time energy sequence, and the vibration signal is segmented at the time point when the energy peak exceeds the dynamic segmentation threshold.

[0020] Preferably, the time-frequency analysis of the denoised vibration signal using synchronous compression transform to obtain a time-frequency matrix includes:

[0021] Perform continuous wavelet transform on the denoised vibration signal to obtain the wavelet coefficients and calculate the instantaneous frequency;

[0022] Redistributing the wavelet coefficients to instantaneous frequency positions by a synchronous compression operator to generate a time-frequency matrix;

[0023] Preferably, the extracting texture features from the time-frequency matrix based on an image processing method includes:

[0024] At least one of the methods of invariant moment, Tamura texture feature and grayscale gradient co-occurrence matrix is ​​used to extract texture features from the time-frequency matrix.

[0025] Preferably, the step of extracting time domain features from the segmented vibration signal comprises:

[0026] The time domain features are obtained by calculating the energy and kurtosis coefficient of each segmented vibration signal in the segmented vibration signal.

[0027] In a second aspect, the present invention further provides a device for extracting characteristics of a vibration-associated signal of an on-load tap changer, comprising:

[0028] A signal acquisition unit is configured to acquire a vibration accompanying signal during operation of the on-load tap changer.

[0029] A signal denoising unit is configured to perform denoising processing on the vibration accompanying signal by using CEEMDAN to obtain a denoised vibration signal.

[0030] A signal segmentation unit is configured to segment the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain a segmented vibration signal.

[0031] A time-domain feature extraction unit is configured to extract time-domain features from the segmented vibration signal.

[0032] A time-frequency analysis unit is configured to perform time-frequency analysis on the denoised vibration signal by using synchronous compression transformation to obtain a time-frequency matrix.

[0033] A texture feature extraction unit is configured to extract texture features representing signal time-frequency energy distribution patterns from the time-frequency matrix based on an image texture analysis method.

[0034] A feature fusion unit is configured to construct a multi-dimensional vibration feature vector according to the time-domain features and the texture features.

[0035] Preferably, the signal denoising unit comprises:

[0036] A component decomposition module is configured to add positive and negative Gaussian white noise to the vibration accompanying signal, perform EMD decomposition, and obtain intrinsic mode components (IMFs) of each order.

[0037] A correlation coefficient calculation module is configured to calculate correlation coefficients of each order of the IMFs and the original signal, and screen out IMF components with correlation coefficients greater than a preset threshold.

[0038] A signal reconstruction module is configured to perform signal reconstruction on the screened-out IMF components to obtain the denoised vibration signal.

[0039] In a third aspect, the present application further provides a computer readable storage medium, which comprises a stored program, wherein the program controls a device where the computer readable storage medium is located to execute the on-load tap changer vibration accompanying signal feature extraction method as described above when the program is running.

[0040] In a fourth aspect, the present application further provides a processor configured to run a program, wherein the program executes the on-load tap changer vibration accompanying signal feature extraction method as described above when the program is running.

[0041] Compared with the prior art, the present application has the following advantages:

[0042] The load tap changer vibration accompanying signal feature extraction method of the present application adopts CEEMDAN to carry out denoising processing on the vibration accompanying signal, can effectively remove the noise component in the signal, and improve the signal-to-noise ratio of the signal; based on short-time energy and dynamic threshold segmentation, the vibration signal after denoising is segmented, different stages can be adaptively divided according to the energy change of the signal, and the key feature segment in the operation process of the load tap changer is accurately divided; the time domain features of the segmented vibration signal are extracted, and the texture features are extracted by using synchronous compression transformation for time-frequency analysis, and the comprehensive extraction of signal features from two dimensions of time domain and frequency domain is realized. The time domain features can reflect the local energy and impact characteristics of the signal, the texture features can represent the mode of signal time-frequency energy distribution, and the multi-dimensional feature extraction method can more comprehensively reflect the running state of the load tap changer. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the specific embodiments or the prior art description will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, each element or part is not necessarily drawn according to the actual proportion.

[0044] Figure 1 The flow chart of the load tap changer vibration accompanying signal feature extraction method of the present application.

[0045] Figure 2 The principle diagram of the load tap changer vibration accompanying signal feature extraction system of the present application. DETAILED DESCRIPTION

[0046] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0047] It should be understood that when used in the present specification, the terms "comprise" and "include" indicate the existence of described features, integers, steps, operations, elements, and / or components, but do not exclude one or more other features, integers, steps, operations, elements, components, and / or sets thereof.

[0048] It should also be understood that the terms used in the present application specification are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the present application specification and the appended claims, unless otherwise clearly indicated by the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0049] It should be further understood that the term "and / or" used in the description of the present application means one or more of the associated listed items and all possible combinations thereof, and includes these combinations.

[0050] The following examples are described with reference to Figure 1 and Figure 2 .

[0051] The embodiment of the present application provides a load-tap-changer vibration accompanying signal feature extraction method, comprising:

[0052] Step S1, acquiring a vibration accompanying signal when a load-tap-changer operates;

[0053] For example, in a certain substation, a vibration sensor is installed near the load-tap-changer, when the load-tap-changer operates, the sensor collects the vibration accompanying signal generated by the load-tap-changer in real time, and converts the analog signal into a digital signal through a data acquisition card and stores it into a computer.

[0054] Step S2, performing denoising processing on the vibration accompanying signal by using complete adaptive noise set empirical mode decomposition (CEEMDAN) to obtain a denoised vibration signal;

[0055] Specifically, the denoising processing on the vibration accompanying signal by using CEEMDAN comprises:

[0056] Adding positive and negative Gaussian white noise to the vibration accompanying signal, and performing EMD decomposition to obtain each order intrinsic mode component IMF;

[0057] In a specific implementation, the collected load-tap-changer vibration accompanying signal is added with positive and negative Gaussian white noise. Then, the noise-added signal is decomposed by EMD to obtain each order intrinsic mode component IMF with different frequencies and representing different vibration characteristics, for example, an IMF component with high frequency reflecting short-term rapid vibration and an IMF component with low frequency reflecting long-term slow vibration. After adding noise to the original signal x(t) to generate a new signal, the new signal is represented as:

[0058]

[0059] Wherein, n i (t) is the i-th Gaussian white noise, and ∈ is a noise intensity coefficient;

[0060] Correlation coefficients of each order IMF and the original signal are calculated, and an IMF component with a correlation coefficient greater than a preset threshold is screened out;

[0061] The correlation coefficient screening condition is:

[0062]

[0063] wherein, p k is the correlation coefficient of the kth IMF and the original signal, p threshold is a preset threshold, IMF k is the kth IMF component;

[0064] By calculating the correlation coefficient of each IMF component obtained by decomposition and the original vibration accompanying signal, the IMF component with a correlation coefficient greater than a preset threshold p threshold is screened out. The screened component contains vibration information in the original signal and eliminates most IMF components corresponding to noise interference.

[0065] The screened IMF component is subjected to signal reconstruction to obtain a denoised vibration signal.

[0066] Step S3, segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain a segmented vibration signal;

[0067] Specifically, the segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation comprises:

[0068] calculating a short-time energy sequence of the denoised vibration signal through a sliding window;

[0069] Taking the denoised on-load tap changer vibration signal as an object, a sliding window with a suitable length is set, the window length is N, and the window is slid along the signal sequence to calculate the short-time energy of the signal in each window in turn to form a short-time energy sequence. The sequence can reflect the trend of the change of the signal energy over time and is helpful to capture the key fluctuation moment of the signal. The energy E n of the nth point in the window is represented as:

[0070]

[0071] wherein, x(k) is the kth sampling point of the denoised signal;

[0072] The OTSU algorithm is used to select a dynamic segmentation threshold for the short-time energy sequence, and the time point at which the energy peak value exceeds the dynamic segmentation threshold is used to segment the vibration signal.

[0073] The OTSU algorithm calculates the inter-class variance of the foreground (energy part peak value) and the background (energy lower part) under each threshold by traversing all possible thresholds, and selects the threshold that maximizes the inter-class variance as the optimal threshold. For example, the short-time energy sequence is sorted to obtain an ordered energy value sequence. All possible thresholds T (each element in the energy value sequence) are traversed, and the short-time energy sequence is divided into two categories, foreground (energy value≥T) and background (energy value

[0074] automatically calculating an optimal threshold T of the energy sequence * , satisfying:

[0075]

[0076] wherein ω1 and ω2 are the proportions of two types of samples, and μ1 and μ2 are the means of two types;

[0077] The threshold calculated by the OTSU algorithm can adaptively adapt to the distribution of signal energy. The time point at which the energy peak in the short-time energy sequence exceeds the dynamic segmentation threshold is taken as a demarcation point. The entire vibration signal is divided into multiple segments, and each segment corresponds to a different state or action phase in the operation of the on-load tap-changer.

[0078] Step S4, extracting time-domain features from the segmented vibration signal;

[0079] Specifically, the extracting time-domain features from the segmented vibration signal comprises:

[0080] The time-domain features are obtained by calculating the energy and kurtosis coefficient of each segmented vibration signal in the segmented vibration signal.

[0081] Segmented energy E s of the s-th segment is represented as:

[0082]

[0083] wherein M s is the number of sampling points of the s-th segment, and x s (m) is the signal value within the segment;

[0084] Kurtosis coefficient K s of the s-th segment is represented as:

[0085]

[0086] wherein μ s is the mean of the segment, and σ s is the standard deviation;

[0087] The segmented energy reflects the intensity of vibration within the segment; and the calculation of the kurtosis coefficient can reflect the probability distribution characteristics of the segmented signal. If the kurtosis coefficient is large, it indicates that there are more impact components in the signal. These time-domain features can comprehensively reflect the vibration characteristics of the on-load tap-changer in different operating phases.

[0088] Step S5, performing time-frequency analysis on the denoised vibration signal by using synchronous compression transformation to obtain a time-frequency matrix;

[0089] Specifically, the time-frequency analysis on the denoised vibration signal is performed by using the synchronous compression transform to obtain a time-frequency matrix, which includes:

[0090] The continuous wavelet transform is performed on the denoised vibration signal to obtain wavelet coefficients and calculate an instantaneous frequency;

[0091] The wavelet coefficients are:

[0092]

[0093] In the formula, W x The wavelet coefficients of the continuous wavelet transform (CWT) (a, b) represent the transform result of the signal x(t) at the scale a and the position b; a is the scale; b is the translation; ψ is the wavelet base function, ψ * is the complex conjugate of ψ;

[0094] The signal is decomposed into different scales and positions by the continuous wavelet transform to obtain a wavelet coefficient matrix, which reflects the characteristics of the signal at different time scales.

[0095] The instantaneous frequency calculation is:

[0096]

[0097] The instantaneous frequency is ω

[0098] The continuous wavelet transform is performed on the denoised on-load tap changer vibration signal to obtain wavelet coefficients and corresponding instantaneous frequencies; the size of the wavelet coefficients reflects the energy intensity of the signal at different scales (corresponding frequencies), and the instantaneous frequency gives the change of the signal frequency with time.

[0099] The wavelet coefficients are redistributed to the instantaneous frequency positions by the synchronous compression operator to generate a time-frequency matrix;

[0100] The time-frequency matrix element is represented as:

[0101]

[0102] In the formula, ω l is the discrete frequency; Δω is the frequency resolution; a -3 / 2 is the adjustment of the scale factor, which is used to compensate the influence of the scale change on the time-frequency matrix;

[0103] By using the synchronous compression operator, the wavelet coefficients are redistributed to the accurate instantaneous frequency positions according to the calculated instantaneous frequencies to generate a clearer time-frequency matrix, which has improved resolution in time and frequency and more detailedly presents the distribution characteristics of the energy of the vibration signal at different times and different frequencies.

[0104] Step S6, extracting texture features representing signal time-frequency energy distribution patterns from the time-frequency matrix based on image texture analysis methods;

[0105] The texture features are extracted from the time-frequency matrix by using at least one of the invariant moments, the Tamura texture features and the gray level co-occurrence matrix.

[0106] The invariant moments are a commonly used image feature extraction method. For a time-frequency matrix (which can be regarded as an image), the seventh-order invariant moment calculation formula is based on the normalized central moment η pq The seventh-order invariant moment is calculated as follows:

[0107] First moment (size feature):

[0108] φ1 = η 20 + η 02

[0109] Wherein, the first moment is used to represent the extension degree of time-frequency energy distribution;

[0110] Second moment (directional independence):

[0111]

[0112] Wherein, the second moment is used to describe the isotropic degree of energy distribution;

[0113] Third moment (skewness feature):

[0114] φ3 = (η 30 - 3η 12 ) 2 + (η 03 - 3η 21 ) 2

[0115] Wherein, the third moment is used to detect the skewness of time-frequency energy distribution;

[0116] Fourth moment (peak feature):

[0117] φ4 = (η 30 + η 12 ) 2 + (η 03 + η 21 ) 2

[0118] Wherein, the fourth moment is used to quantify the peak degree of energy concentration area;

[0119] Fifth moment (complex deformation):

[0120] φ5 = (η 30 - 3η 12)(η 30 +η 12 )[(η 30 +η 12 ) 2 -3(η 03 +η 21 ) 2 ]

[0121] where the fifth-order moment is used to identify local skewness in the energy distribution;

[0122] sixth-order moment (asymmetry)

[0123] φ6= (η 20 -η 02 )[(η 30 +η 12 ) 2 -(η 03 +η 21 ) 2 ]+4η 11 (η 30 +η 12 )(η 03 +η 21 )

[0124] where the sixth-order moment is used to characterize mirror asymmetry in the time- frequency structure;

[0125] seventh-order moment (high-order distortion)

[0126] φ7= (3η 21 -η 03 )(η 30 +η 12 )[(η 30 +η 12 ) 2 -3(η 03 +η 21 ) 2 ]

[0127] +(3η 12 -η 30 )(η 21 +η 03 )[3(η 30 +η 12 ) 2 -(η 03 +η 21 ) 2 ]

[0128] where the seventh-order moment is used to detect subtle distortion patterns in the energy distribution.

[0129] The first to seventh order invariant moments are numbered, and by randomly selecting a group of vibration accompanying signals, the invariant moment texture features are extracted from the time-frequency matrix.

[0130] In step S7, a multi-dimensional vibration feature vector is constructed according to the time domain feature and the texture feature.

[0131] By combining the extracted time domain features (energy, kurtosis coefficient) and texture features together, a multi-dimensional vibration feature vector is constructed, which comprehensively reflects the characteristics of the on-load tap changer vibration accompanying signal in the time domain and the frequency domain.

[0132] The on-load tap changer vibration accompanying signal feature extraction method of the present application adopts CEEMDAN to denoise the vibration accompanying signal, which can effectively remove the noise components in the signal and improve the signal-to-noise ratio; based on short-time energy and dynamic threshold segmentation, the vibration signal after denoising is segmented, which can adaptively divide different stages according to the energy change of the signal, and accurately to the key feature segment in the on-load tap changer operation process; the time domain features of the segmented vibration signal are extracted, and the texture features are extracted by using synchronous compression transformation for time-frequency analysis, which realizes the comprehensive extraction of signal features from two dimensions of time domain and frequency domain. The time domain features can reflect the local energy and impact characteristics of the signal, and the texture features can represent the mode of signal time-frequency energy distribution, and the multi-dimensional feature extraction method more comprehensively reflects the running state of the on-load tap changer.

[0133] The present application also provides an on-load tap changer vibration accompanying signal feature extraction device, comprising:

[0134] A signal acquisition unit is configured to acquire the vibration accompanying signal during the operation of the on-load tap changer.

[0135] A signal denoising unit is configured to use CEEMDAN to denoise the vibration accompanying signal to obtain a denoised vibration signal.

[0136] A signal segmentation unit is configured to segment the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain a segmented vibration signal.

[0137] A time domain feature extraction unit is configured to extract time domain features from the segmented vibration signal.

[0138] A time-frequency analysis unit is configured to use synchronous compression transformation to perform time-frequency analysis on the denoised vibration signal to obtain a time-frequency matrix.

[0139] A texture feature extraction unit is configured to extract texture features representing the mode of signal time-frequency energy distribution from the time-frequency matrix based on an image texture analysis method.

[0140] A feature fusion unit is configured to construct a multi-dimensional vibration feature vector according to the time-domain features and the texture features.

[0141] Specifically, the signal denoising unit comprises:

[0142] A component decomposition module is configured to add positive and negative Gaussian white noises to the vibration accompanying signal, perform EMD decomposition, and obtain each order intrinsic mode component (IMF);

[0143] A correlation coefficient calculation module is configured to calculate correlation coefficients of each order IMF and the original signal, and screen out the IMF components with correlation coefficients greater than a preset threshold.

[0144] A signal reconstruction module is configured to perform signal reconstruction on the screened IMFs, and obtain the denoised vibration signal.

[0145] The functions of each unit in the embodiment are explained as the same as the method for extracting a vibration accompanying signal feature of an on-load tap changer, and the technical effects are the same, which will not be repeated here.

[0146] The embodiment of the application further provides a computer readable storage medium comprising a stored program, wherein the program controls the device where the computer readable storage medium is located to execute the method for extracting a vibration accompanying signal feature of an on-load tap changer when the program is run.

[0147] The technical effects of the embodiment are the same as those of the method for extracting a vibration accompanying signal feature of an on-load tap changer, which will not be repeated here.

[0148] The application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like.

[0149] The embodiment of the application further provides a processor for running a program, wherein the program executes the method for extracting a vibration accompanying signal feature of an on-load tap changer when the program is run.

[0150] The technical effects of the embodiment are the same as those of the method for extracting a vibration accompanying signal feature of an on-load tap changer, which will not be repeated here.

[0151] The processor in the embodiment can be a central processing unit (CPU), a controller, a microcontroller, or other data processing chips.

[0152] Those skilled in the art can appreciate that the units of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components of the examples have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0153] In the embodiments provided by the present application, it should be understood that the division of units is only a logical functional division, and there can be another division manner in actual implementation, for example, multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored, etc.

[0154] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0155] When the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0156] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should be covered in the scope of the description of the present application.

Claims

1. A method for extracting characteristics of vibration signals of an on-load tap changer, characterized in that: include: Obtain vibration signals during operation of on-load tap changer; Using CEEMDAN to denoise the vibration signal to obtain a denoised vibration signal; Segmenting the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain segmented vibration signals; extracting time domain features from the segmented vibration signal; Performing time-frequency analysis on the denoised vibration signal using synchronous compression transform to obtain a time-frequency matrix; Extracting texture features representing the time-frequency energy distribution pattern of the signal from the time-frequency matrix based on an image texture analysis method; A multi-dimensional vibration feature vector is constructed according to the time domain feature and the texture feature.

2. The method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to claim 1, characterized in that: The denoising process of the vibration associated signal by using CEEMDAN includes: Add positive and negative Gaussian white noise to the vibration accompanying signal, perform EMD decomposition, and obtain the intrinsic modal components (IMFs) of each order; Calculate the correlation coefficient between each order IMF and the original signal, and filter out the IMF components whose correlation coefficient is greater than the preset threshold; The filtered IMF components are reconstructed to obtain the denoised vibration signal.

3. The method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to claim 1, characterized in that: The segmenting of the denoised vibration signal based on short-time energy and dynamic threshold segmentation includes: Calculating a short-time energy sequence of the denoised vibration signal through a sliding window; The OTSU algorithm is used to select a dynamic segmentation threshold for the short-time energy sequence, and the vibration signal is segmented at the time point when the energy peak exceeds the dynamic segmentation threshold.

4. The method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to claim 3, characterized in that: The time-frequency analysis of the denoised vibration signal is performed by using synchronous compression transformation to obtain a time-frequency matrix including: Perform continuous wavelet transform on the denoised vibration signal to obtain the wavelet coefficients and calculate the instantaneous frequency; The wavelet coefficients are redistributed to instantaneous frequency positions through a synchronous compression operator to generate a time-frequency matrix.

5. The method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to claim 1, characterized in that: The extracting of texture features from the time-frequency matrix based on an image processing method comprises: At least one of the methods of invariant moment, Tamura texture feature and grayscale gradient co-occurrence matrix is ​​used to extract texture features from the time-frequency matrix.

6. The method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to claim 1, characterized in that: The extracting time domain features of the segmented vibration signal comprises: The time domain features are obtained by calculating the energy and kurtosis coefficient of each segmented vibration signal in the segmented vibration signal.

7. A device for extracting characteristics of vibration signals of an on-load tap changer, characterized in that: A method for extracting features of a vibration-associated signal of an on-load tap changer as claimed in any one of claims 1 to 6, comprising: A signal acquisition unit is used to obtain vibration signals accompanying the operation of the on-load tap changer; A signal denoising unit, configured to perform denoising processing on the vibration accompanying signal using CEEMDAN to obtain a denoised vibration signal; a signal segmentation unit, configured to segment the denoised vibration signal based on short-time energy and dynamic threshold segmentation to obtain segmented vibration signals; A time domain feature extraction unit, configured to extract time domain features from the segmented vibration signal; A time-frequency analysis unit, configured to perform time-frequency analysis on the denoised vibration signal using synchronous compression transform to obtain a time-frequency matrix; A texture feature extraction unit, configured to extract texture features representing a time-frequency energy distribution pattern of a signal from the time-frequency matrix based on an image texture analysis method; A feature fusion unit is used to construct a multi-dimensional vibration feature vector according to the time domain feature and the texture feature.

8. The device for extracting characteristics of vibration-associated signals of an on-load tap changer according to claim 7, characterized in that: The signal denoising unit includes: A component decomposition module is used to add positive and negative Gaussian white noise to the vibration accompanying signal, perform EMD decomposition, and obtain the intrinsic modal components (IMFs) of each order; The correlation coefficient calculation module is used to calculate the correlation coefficient between each order IMF and the original signal, and filter out the IMF components with correlation coefficients greater than a preset threshold; The signal reconstruction module is used to reconstruct the filtered IMF components to obtain the denoised vibration signal.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for extracting characteristics of vibration accompanying signals of an on-load tap changer according to any one of claims 1 to 6.

10. A processor, characterized in that: The processor is used to run a program, wherein the program, when running, executes the method for extracting features of vibration accompanying signals of an on-load tap changer according to any one of claims 1 to 6.

Images