A method and device for monitoring transformer winding status under short circuit impact

Abstract

The embodiment of the present invention discloses a method and device for monitoring the state of a transformer winding under a short-circuit impact. The method comprises: acquiring a short-circuit impact current and a short-circuit impact vibration signal of a transformer when the transformer is subjected to a short-circuit impact, obtaining a time domain envelope vector under the short-circuit impact and a vibration signal characteristic vector under the short-circuit impact, using the eigenvalue and eigenvector of a mutual information matrix constructed according to the time domain envelope vector and the vibration signal characteristic vector, calculating a weighted fusion similarity of the cosine similarity of the characteristic vector, comparing the weighted fusion similarity with a control limit interval, thereby determining the state of the transformer winding, and realizing online monitoring of the transformer winding state under the short-circuit impact.

Description

Technical Field

[0001] The present invention relates to the technical field of signal monitoring, and in particular to a method and a device for monitoring the state of transformer windings under short-circuit impact. Background Art

[0002] Winding deformation detection of power transformers is one of the conventional test items of transformers at present. The commonly used detection methods are mainly short-circuit impedance method and frequency response analysis method. The former mainly indirectly judges whether the winding is deformed based on the change of the transformer winding leakage reactance. This method has low sensitivity and fault detection rate. It can only give a relatively accurate diagnosis result when the overall deformation of the transformer coil is serious, and it is powerless against loose transformer windings. The latter regards the transformer winding as a distributed parameter network, and describes its characteristics in the frequency domain by the transfer function according to the change of the corresponding distributed parameters when the transformer winding is loose or deformed, and then judges the mechanical state of the winding. However, the frequency response waveform of this method is relatively complex, especially the high-frequency parameters are easily affected by related interference, and rich experience is required to accurately judge the winding state. Therefore, how to efficiently and accurately monitor the state of transformer windings under short-circuit impact needs to be improved. Summary of the invention

[0003] The main purpose of the present invention is to provide a method and device for monitoring the state of transformer windings under short-circuit impact, which can solve the problem of low sensitivity and accuracy in monitoring the state of transformer windings under short-circuit impact in the prior art.

[0004] To achieve the above object, the present invention provides a method for monitoring transformer winding status under short circuit impact in a first aspect, the method comprising:

[0005] Collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, and calculating the time domain envelope vector under the short-circuit impact by using the short-circuit impact current of the transformer based on Hilbert transform, and calculating the vibration signal characteristic vector under the short-circuit impact by using the short-circuit impact vibration signal of the transformer based on adaptive discrete transform;

[0006] Constructing a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculating the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue;

[0007] A weighted fusion similarity of the cosine similarity of the feature vector is obtained according to the cosine similarity of the feature vector, the weighted fusion similarity is compared with the control limit interval, and the transformer winding state is determined according to the comparison result.

[0008] In combination with the first aspect, in a possible implementation method, the above-mentioned method is based on adaptive discrete transformation and uses the transformer short-circuit impact vibration signal to calculate the vibration signal characteristic vector under short-circuit impact, including: calculating the vibration signal spectrum distribution according to the transformer short-circuit impact vibration signal, and normalizing the vibration signal spectrum distribution; calculating an adaptive window function according to the normalized vibration signal spectrum distribution, and performing an adaptive discrete transformation on the transformer short-circuit impact vibration signal based on the adaptive window function to obtain a vibration signal transformation spectrum matrix; and obtaining a vibration signal characteristic vector according to the vibration signal transformation spectrum matrix.

[0009] In combination with the first aspect, in one possible implementation method, the above-mentioned obtaining the vibration signal feature vector according to the vibration signal transformation spectrum matrix includes: performing matrix transformation on the vibration signal transformation spectrum matrix to obtain a transformation matrix, performing singular value decomposition on the transformation matrix to obtain a diagonal matrix; extracting the elements on the diagonal of the diagonal matrix in sequence, and constructing the vibration signal feature vector using the extracted elements.

[0010] In combination with the first aspect, in one possible implementation method, the above-mentioned calculation of the cosine similarity of the feature vector based on the eigenvalues ​​and the eigenvector includes: calculating the cumulative contribution rate of the eigenvalues; selecting the eigenvectors corresponding to the first z eigenvalues ​​whose cumulative contribution rate is equal to a preset threshold, where z is a positive integer; and calculating the cosine similarity of the eigenvectors corresponding to the first z eigenvalues.

[0011] In combination with the first aspect, in one possible implementation method, the above-mentioned determination of the transformer winding state based on the comparison result includes: when the comparison result is that the weighted fusion similarity is not within the control limit interval, it is judged that the transformer winding has been deformed; when the comparison result is that the weighted fusion similarity is within the control limit interval, it is judged that the transformer winding has not been deformed.

[0012] In combination with the first aspect, in a possible implementation manner, the above Hilbert transform formula is as follows:

[0013]

[0014]

[0015] Among them, i(t) represents the transformer short-circuit impulse current, b(t) represents the time domain envelope, t represents time, and j represents the imaginary unit. It means that the convolution meets the requirement. is the Hilbert transform result of i(t).

[0016] In combination with the first aspect, in a possible implementation manner, the above-mentioned calculation of the adaptive window function according to the normalized vibration signal spectrum distribution includes:

[0017] The calculation formula is:

[0018]

[0019]

[0020] Wherein, w(m,c(n)) represents the adaptive window function, m represents the variable used in the adaptive discrete transform, the value of m ranges from 0 to N-1, N is the length of the vibration signal; c(n) is the window length at each frequency in the spectral component of the vibration signal; X' v (n) represents the normalized frequency distribution of the vibration signal.

[0021] In combination with the first aspect, in a possible implementation manner, the formula for calculating the cosine similarity of the feature vector is as follows:

[0022]

[0023] Where: CS(d i ,d j ) represents the cosine similarity of feature vectors; T represents transposition; represents the 2-norm of the eigenvector; d i ,d j represents the eigenvector; i and j represent the ordinal number of the eigenvalue.

[0024] In combination with the first aspect, in a possible implementation, the calculation formula for the weighted fusion similarity of the cosine similarity of the feature vector obtained above is:

[0025]

[0026] Among them, χ represents the weighted fusion similarity; λ i , j , k represents the eigenvalue, where λ i , j , k denote the i-th eigenvalue, j-th eigenvalue, and k-th eigenvalue respectively; CS(d i , d j ) represents the cosine similarity of feature vectors.

[0027] To achieve the above object, the second aspect of the present invention provides a transformer winding state monitoring device under short circuit impact, the device comprising:

[0028] Signal processing module: used for collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, and based on Hilbert transform, using the short-circuit impact current of the transformer to calculate the time domain envelope vector under the short-circuit impact, and based on adaptive discrete transform, using the short-circuit impact vibration signal of the transformer to calculate the vibration signal feature vector under the short-circuit impact;

[0029] Similarity calculation module: used to construct a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculate the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue;

[0030] State determination module: used for obtaining weighted fusion similarity of cosine similarity of feature vectors according to cosine similarity of feature vectors, comparing the weighted fusion similarity with control limit interval, and determining transformer winding state according to comparison result.

[0031] The embodiment of the present invention has the following beneficial effects: by collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, the time domain envelope vector under the short-circuit impact and the vibration signal characteristic vector under the short-circuit impact are obtained, and the eigenvalue and eigenvector of the mutual information matrix constructed according to the time domain envelope vector and the vibration signal characteristic vector are used to calculate the weighted fusion similarity of the cosine similarity of the characteristic vector, and the weighted fusion similarity is compared with the control limit interval, so as to determine the transformer winding state. In this technical solution, since the transformer short-circuit impact vibration signal is very easy to monitor online, the time domain envelope vector and the vibration signal characteristic vector are obtained by the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, while ensuring the timeliness of the monitoring of the transformer short-circuit impact current and the transformer short-circuit impact vibration signal, the accuracy of the obtained time domain envelope vector and the vibration signal characteristic vector under the transformer short-circuit impact is improved. In addition, after calculating the cosine similarity of the eigenvalue and the eigenvector, the weighted fusion similarity is also calculated based on the cosine similarity of the eigenvector, which improves the accuracy of the transformer winding state judgment. Moreover, the transformer winding state is directly judged by comparing the weighted fusion similarity with the control limit interval, which is efficient and simple. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0033] in:

[0034] Figure 1 A schematic flow chart of a method for monitoring transformer winding status under short-circuit impact in an embodiment of the present invention;

[0035] Figure 2 It is a structural block diagram of a transformer winding state monitoring device under short-circuit impact in an embodiment of the present invention;

[0036] Figure 3 4 is a structural block diagram of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] The embodiment of the present invention provides a method for monitoring the state of transformer windings under short circuit impact, referring to Figure 1 , Figure 1 A flow chart of a method for monitoring transformer winding status under short-circuit impact provided by an embodiment of the present invention is shown as follows: Figure 1 As shown, the method specifically comprises the following steps:

[0039] Step S101, collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, based on the Hilbert transform, using the short-circuit impact current of the transformer to calculate the time domain envelope vector under the short-circuit impact, and based on the adaptive discrete transform, using the short-circuit impact vibration signal of the transformer to calculate the vibration signal feature vector under the short-circuit impact.

[0040] Step S102: construct a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculate the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue.

[0041] Step S103: obtaining a weighted fusion similarity of the cosine similarity of the feature vector according to the cosine similarity of the feature vector, comparing the weighted fusion similarity with the control limit interval, and determining the transformer winding state according to the comparison result.

[0042] Since the transformer winding is mainly wound by electromagnetic wire according to a certain structure, if it is regarded as a mechanical structure composed of several natural frequencies and corresponding modes, any change in the winding structure can be reflected from the change in its mechanical characteristics. Under short-circuit impact, the vibration of the transformer box wall is mainly composed of the vibration of the winding under the action of the short-circuit current. As long as the mechanical characteristics of the transformer winding change, it can be reflected from its mechanical vibration response characteristics. Therefore, this embodiment detects the state of the transformer winding under short-circuit impact by online monitoring of the transformer short-circuit impact vibration signal. Specifically, step S101 is executed.

[0043] The transformer short-circuit impact current and the transformer short-circuit impact vibration signal are collected when the transformer is subjected to a short-circuit impact. In this embodiment, the transformer short-circuit impact vibration signal can be obtained by a vibration acceleration sensor placed on the wall of the transformer oil tank, wherein the sampling frequency of the transformer short-circuit impact current, the transformer short-circuit impact vibration signal and the sound signal is f0, and the transformer short-circuit impact current duration is T s In this embodiment, f0 = 51.2kHz, T s =0.5s.

[0044] It is impossible to intuitively judge the state of the transformer winding under short-circuit impact by directly obtaining the transformer short-circuit impact current and the transformer short-circuit impact vibration signal. In this embodiment, a parameter value is obtained by calculating and processing the transformer short-circuit impact current and the transformer short-circuit impact vibration signal to judge the state of the transformer winding.

[0045] Specifically, after the transformer short-circuit impact current and the transformer short-circuit impact vibration signal are collected, the transformer short-circuit impact current is used to calculate the time domain envelope vector under the short-circuit impact, and the transformer short-circuit impact vibration signal is used to calculate the vibration signal feature vector under the short-circuit impact.

[0046] In this embodiment, the time domain envelope vector under the short circuit impact is calculated based on Hilbert transform, wherein the Hilbert transform calculation formula is:

[0047]

[0048]

[0049] In the formula, i(t) represents the transformer short-circuit impulse current, b(t) represents the time domain envelope, t represents time, and j represents the imaginary unit. It means that the convolution meets the requirement. is the Hilbert transform result of i(t).

[0050] After the time domain envelope under the short circuit impact is calculated, the time domain envelope is represented as a length of 1×D i dimensional row vector, and obtain the time domain envelope vector, where D i is the length of the time domain envelope.

[0051] In this embodiment, based on the adaptive discrete transformation, the vibration signal characteristic vector under the short-circuit impact is calculated using the transformer short-circuit impact vibration signal, and the specific steps are as follows:

[0052] Step S1011, calculating the frequency spectrum distribution of the vibration signal according to the transformer short-circuit impact vibration signal, and performing normalization processing on the frequency spectrum distribution.

[0053] Step S1012, calculating an adaptive window function according to the normalized frequency spectrum distribution of the vibration signal, and performing an adaptive discrete transformation on the transformer short-circuit impact vibration signal based on the adaptive window function to obtain a vibration signal transformation spectrum matrix.

[0054] Step S1013: Obtain a vibration signal feature vector according to the vibration signal transformation spectrum matrix.

[0055] The vibration signal spectrum distribution of the transformer short-circuit impact vibration signal is calculated based on Fourier transform, and the vibration signal spectrum distribution is normalized. The calculation formula is:

[0056]

[0057] Where: X v (n) represents the frequency distribution of the vibration signal before normalization, X′ v (n) represents the normalized frequency distribution of the vibration signal.

[0058] The adaptive window function is calculated according to the normalized vibration signal spectrum distribution, and its calculation formula is:

[0059]

[0060]

[0061] Wherein, w(m,c(n)) represents the adaptive window function, m represents the variable used in the adaptive discrete transform, the value of m ranges from 0 to N-1, N is the length of the vibration signal; c(n) is the window length at each frequency in the spectrum distribution of the vibration signal; X' v (n) represents the normalized frequency distribution of the vibration signal.

[0062] Based on the adaptive window function, the transformer short-circuit impact vibration signal is adaptively discretely transformed to obtain the vibration signal transformation spectrum matrix, and its calculation formula is:

[0063]

[0064] Wherein, T is the sampling interval, l(n) is the window length of the adaptive window function, i is the imaginary unit, m represents the variable used in the adaptive discrete transform, the value of m is from 0 to N-1, the value range of n is 0, 1, 2, ..., N-1, k = 0, 1, 2, ..., N-1, and N is the length of the vibration signal.

[0065] After the vibration signal transformation spectrum matrix is ​​obtained, the vibration signal feature vector is obtained according to the vibration signal transformation spectrum matrix. The specific steps are as follows:

[0066] Step S10131, performing matrix transformation on the vibration signal transformation spectrum matrix to obtain a transformation matrix, and performing singular value decomposition on the transformation matrix to obtain a diagonal matrix.

[0067] Step S10132: extract the elements on the diagonal of the diagonal matrix in sequence, and construct a vibration signal feature vector using the extracted elements.

[0068] Among them, the vibration signal transformation spectrum matrix is ​​transformed into a transformation matrix, and the transformation matrix is ​​decomposed into a singular value to obtain a diagonal matrix. The calculation formula is:

[0069] E=W T W

[0070] E=QΛQ -1

[0071] In the formula: W represents the vibration signal transformation spectrum matrix; E represents the transformation matrix obtained by matrix transformation of the vibration signal transformation spectrum matrix; Q represents the singular matrix of the transformation matrix; Λ represents the diagonal matrix of the transformation matrix; T represents transpose.

[0072] Then, the elements on the diagonal of the diagonal matrix are extracted in sequence, and the vibration signal feature vector B is constructed using the extracted elements. v , the vibration signal feature vector is 1×D v dimensional row vector, where D v is the number of elements on the diagonal of the diagonal matrix Λ.

[0073] Since the transformer short-circuit impact vibration signal is very easy to monitor online, the time domain envelope vector and the vibration signal characteristic vector are obtained through the transformer short-circuit impact current and the transformer short-circuit impact vibration signal when the transformer is subjected to a short-circuit impact. This ensures the timeliness of the monitoring of the transformer short-circuit impact current and the transformer short-circuit impact vibration signal, and improves the accuracy of the obtained time domain envelope vector and the vibration signal characteristic vector under the transformer short-circuit impact.

[0074] After the time domain envelope vector and the vibration signal eigenvector are obtained, a mutual information matrix is ​​constructed based on the time domain envelope vector and the vibration signal eigenvector, and the cosine similarity of the eigenvector is calculated according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue, that is, step S102 is executed.

[0075] In this embodiment, a p×p-dimensional mutual information matrix M is constructed based on the time domain envelope vector of the transformer short-circuit impulse current and the vibration signal feature vector, where the dimension p=min(D i ,D v ), the mutual information matrix M can be expressed as

[0076]

[0077] Here, m ij (i, j = 1, 2, ..., p) represents the mutual information value between the i-th element in the time domain envelope vector and the j-th element of the vibration signal feature vector.

[0078] The Jacobi method is used to calculate the eigenvalues ​​and eigenvectors of the mutual information matrix, and the q eigenvalues ​​of the mutual information matrix calculated are recorded as λ1, λ2, ..., λ q , and satisfies λ1≥λ2≥……≥λ q ≥0, where the eigenvector corresponding to the i-th eigenvalue is d i =(d i1 , d i2 , …, d iq ) T , i = 1, 2, ..., q. The Jacobi method in this step is a commonly used mathematical method in the art, and therefore, the inventors will not describe it in detail here.

[0079] After the eigenvalues ​​and eigenvectors of the mutual information matrix are calculated, the cosine similarity of the eigenvectors is calculated according to the eigenvalues ​​of the mutual information matrix and the eigenvectors of the eigenvalues. The specific steps are as follows:

[0080] Step S1021, calculating the cumulative contribution rate of the eigenvalue.

[0081] Step S1022: Select the eigenvectors corresponding to the first z eigenvalues ​​whose cumulative contribution rates are equal to the preset threshold, where z is a positive integer.

[0082] Step S1023: Calculate the cosine similarity of the eigenvectors corresponding to the first z eigenvalues.

[0083] The calculation formula for calculating the cumulative contribution rate α of the eigenvalue is:

[0084]

[0085] In the formula, α represents the cumulative contribution rate; λ i , q represent the i-th eigenvalue and the q-th eigenvalue respectively.

[0086] Set the preset threshold of the cumulative contribution rate according to the requirements, and select the eigenvectors d1, d2, ..., d corresponding to the first z eigenvalues ​​whose cumulative contribution rates are equal to the preset threshold. z , and calculate its cosine similarity, where the formula for calculating the cosine similarity of the feature vector is as follows:

[0087]

[0088] Where: CS(d i ,d j ) represents the cosine similarity of feature vectors; T represents transposition; represents the 2-norm of the eigenvector; d i ,d j represents the eigenvector; i and j represent the ordinal number of the eigenvalue.

[0089] In this embodiment, the preset threshold is set to 0.99, z=3, and the eigenvectors d1, d2, and d3 corresponding to the first three eigenvalues ​​corresponding to the cumulative contribution rate α=0.99 are selected, and their cosine similarity is calculated.

[0090] After the cosine similarity is obtained, a weighted fusion similarity of the cosine similarity of the feature vector is obtained according to the cosine similarity of the feature vector, the weighted fusion similarity is compared with the control limit interval, and the transformer winding state is determined according to the comparison result, that is, step S103 is executed.

[0091] Among them, the calculation formula of the weighted fusion similarity of the cosine similarity of the feature vector is:

[0092]

[0093] Among them, χ represents the weighted fusion similarity; λ i , j , k represents the eigenvalue, where λ i , j , k denote the i-th eigenvalue, j-th eigenvalue, and k-th eigenvalue respectively; CS(d i , d j ) represents the cosine similarity of feature vectors.

[0094] In this embodiment, the control limit interval Ψ is determined according to the 3σ criterion, and the weighted fusion similarity χ is compared with the control limit interval Ψ. When the weighted fusion similarity is not within the control limit interval, it is judged that the transformer winding has been deformed; when the weighted fusion similarity is within the control limit interval, it is judged that the transformer winding has not been deformed.

[0095] After calculating the cosine similarity of the eigenvalue and the eigenvector, the weighted fusion similarity is also calculated based on the cosine similarity of the eigenvector, which improves the accuracy of transformer winding state judgment. Moreover, the transformer winding state is directly judged by comparing the weighted fusion similarity with the control limit interval, which is efficient and simple.

[0096] The above describes a method for monitoring the state of a transformer winding under a short-circuit impact provided by an embodiment of the present invention. In order to better implement the above method, the following describes a device 20 for monitoring the state of a transformer winding under a short-circuit impact provided by an embodiment of the present invention. Figure 2 As shown, the device 20 comprises:

[0097] Signal processing module 201: used for collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, and based on Hilbert transform, using the short-circuit impact current of the transformer to calculate the time domain envelope vector under the short-circuit impact, and based on adaptive discrete transform, using the short-circuit impact vibration signal of the transformer to calculate the vibration signal feature vector under the short-circuit impact;

[0098] Similarity calculation module 202: used to construct a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculate the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue;

[0099] The state determination module 203 is used to obtain the weighted fusion similarity of the cosine similarity of the feature vector according to the cosine similarity of the feature vector, compare the weighted fusion similarity with the control limit interval, and determine the transformer winding state according to the comparison result.

[0100] In a possible implementation, the signal processing module 201 is specifically used to: calculate the frequency spectrum distribution of the vibration signal according to the transformer short-circuit impact vibration signal, and normalize the frequency spectrum distribution of the vibration signal; calculate the adaptive window function according to the normalized frequency spectrum distribution of the vibration signal, and perform adaptive discrete transformation on the transformer short-circuit impact vibration signal based on the adaptive window function to obtain a vibration signal transformation spectrum matrix;

[0101] The vibration signal characteristic vector is obtained according to the vibration signal transformation spectrum matrix.

[0102] In one possible implementation, the signal processing module 201 is specifically used to: perform matrix transformation on the vibration signal transformation spectrum matrix to obtain a transformation matrix, perform singular value decomposition on the transformation matrix to obtain a diagonal matrix; sequentially extract the elements on the diagonal of the diagonal matrix, and use the extracted elements to construct a vibration signal feature vector.

[0103] In one possible implementation, the similarity calculation module 202 is specifically used to: calculate the cumulative contribution rate of the eigenvalues; select the eigenvectors corresponding to the first z eigenvalues ​​whose cumulative contribution rate is equal to a preset threshold, where z is a positive integer; and calculate the cosine similarity of the eigenvectors corresponding to the first z eigenvalues.

[0104] In one possible implementation, the state determination module 203 is specifically used to: when the comparison result is that the weighted fusion similarity is not within the control limit interval, it is judged that the transformer winding has been deformed; when the comparison result is that the weighted fusion similarity is within the control limit interval, it is judged that the transformer winding has not been deformed.

[0105] In the above-mentioned device, by collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, the time domain envelope vector under the short-circuit impact and the vibration signal characteristic vector under the short-circuit impact are obtained, and the eigenvalue and eigenvector of the mutual information matrix constructed according to the time domain envelope vector and the vibration signal characteristic vector are used to calculate the weighted fusion similarity of the cosine similarity of the characteristic vector, and the weighted fusion similarity is compared with the control limit interval to determine the transformer winding state. In this technical solution, since the transformer short-circuit impact vibration signal is very easy to monitor online, the time domain envelope vector and the vibration signal characteristic vector are obtained by the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, which ensures the timeliness of the monitoring of the transformer short-circuit impact current and the transformer short-circuit impact vibration signal, and improves the accuracy of the acquired time domain envelope vector and the vibration signal characteristic vector under the transformer short-circuit impact. In addition, after calculating the cosine similarity of the eigenvalue and the eigenvector, the weighted fusion similarity is also calculated based on the cosine similarity of the eigenvector, which improves the accuracy of the transformer winding state judgment. Moreover, the transformer winding state is directly judged by comparing the weighted fusion similarity with the control limit interval, which is efficient and simple.

[0106] Figure 3 FIG. 1 shows an internal structure diagram of a computer device in an embodiment. The computer device may be a terminal or a server. Figure 3As shown, the computer device includes a processor, a memory and a network interface connected via a system bus. The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor may implement each step in the above method embodiment. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor may implement each step in the above method embodiment. Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0107] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0108] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, 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.

[0109] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for monitoring transformer winding status under short-circuit impact, characterized in that: The method comprises: Collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, and calculating the time domain envelope vector under the short-circuit impact by using the short-circuit impact current of the transformer based on Hilbert transform, and calculating the vibration signal characteristic vector under the short-circuit impact by using the short-circuit impact vibration signal of the transformer based on adaptive discrete transform; Constructing a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculating the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue; Obtaining a weighted fusion similarity of the cosine similarity of the feature vector according to the cosine similarity of the feature vector, comparing the weighted fusion similarity with the control limit interval, and determining the transformer winding state according to the comparison result; The calculation formula of the weighted fusion similarity of the cosine similarity of the feature vector is: Among them, χ represents the weighted fusion similarity; λ i , j , k represents the eigenvalue, where λ i , j , k denote the i-th eigenvalue, j-th eigenvalue, and k-th eigenvalue respectively; CS(d i , d j ) represents the cosine similarity of feature vectors.

2. The method according to claim 1, characterized in that The method of calculating the vibration signal characteristic vector under the short-circuit impact by using the transformer short-circuit impact vibration signal based on the adaptive discrete transformation includes: Calculating the frequency spectrum distribution of the vibration signal according to the transformer short-circuit impact vibration signal, and normalizing the frequency spectrum distribution of the vibration signal; Calculating an adaptive window function according to the normalized frequency spectrum distribution of the vibration signal, and performing an adaptive discrete transformation on the transformer short-circuit impact vibration signal based on the adaptive window function to obtain a vibration signal transformation spectrum matrix; The vibration signal characteristic vector is obtained according to the vibration signal transformation spectrum matrix.

3. The method according to claim 2, characterized in that The step of obtaining a vibration signal feature vector according to the vibration signal transformation spectrum matrix comprises: Performing matrix transformation on the vibration signal transformation spectrum matrix to obtain a transformation matrix, and performing singular value decomposition on the transformation matrix to obtain a diagonal matrix; The elements on the diagonal of the diagonal matrix are extracted in sequence, and the vibration signal feature vector is constructed using the extracted elements.

4. The method according to claim 1, characterized in that: The calculating the cosine similarity of the feature vector according to the feature value and the feature vector comprises: Calculating the cumulative contribution rate of the eigenvalue; Select the eigenvectors corresponding to the first z eigenvalues ​​whose cumulative contribution rate is equal to the preset threshold, where z is a positive integer; The cosine similarity of the eigenvectors corresponding to the first z eigenvalues ​​is calculated.

5. The method according to claim 1, characterized in that Determining the transformer winding state according to the comparison result includes: When the comparison result is that the weighted fusion similarity is not within the control limit interval, it is determined that the transformer winding has been deformed; When the comparison result is that the weighted fusion similarity is within the control limit interval, it is determined that the transformer winding is not deformed.

6. The method according to claim 1, characterized in that: The Hilbert transform formula is as follows: Among them, i(t) represents the transformer short-circuit impulse current, b(t) represents the time domain envelope, t represents time, and j represents the imaginary unit. It means that the convolution meets the requirement. is the Hilbert transform result of i(t).

7. The method according to claim 2, characterized in that: The calculation formula for calculating the adaptive window function based on the normalized vibration signal spectrum distribution is: Wherein, w(m,c(n)) represents the adaptive window function, m represents the variable used in the adaptive discrete transform, the value of m ranges from 0 to N-1, N is the length of the vibration signal; c(n) is the window length at each frequency in the spectrum distribution of the vibration signal; X' v (n) represents the normalized frequency distribution of the vibration signal.

8. The method according to claim 4, characterized in that: The formula for calculating the cosine similarity of the feature vector is as follows: Where: CS(d i ,d j ) represents the cosine similarity of feature vectors; T represents transposition; represents the 2-norm of the eigenvector; d i d j represents the eigenvector; i and j represent the ordinal number of the eigenvalue.

9. A transformer winding status monitoring device under short circuit impact, characterized in that: The device comprises: Signal processing module: used for collecting the short-circuit impact current and the short-circuit impact vibration signal of the transformer when it is subjected to a short-circuit impact, and based on Hilbert transform, using the short-circuit impact current of the transformer to calculate the time domain envelope vector under the short-circuit impact, and based on adaptive discrete transform, using the short-circuit impact vibration signal of the transformer to calculate the vibration signal feature vector under the short-circuit impact; Similarity calculation module: used to construct a mutual information matrix according to the time domain envelope vector and the vibration signal eigenvector, and calculate the cosine similarity of the eigenvector according to the eigenvalue of the mutual information matrix and the eigenvector of the eigenvalue; State determination module: used to obtain the weighted fusion similarity of the cosine similarity of the feature vector according to the cosine similarity of the feature vector, compare the weighted fusion similarity with the control limit interval, and determine the transformer winding state according to the comparison result; the calculation formula for obtaining the weighted fusion similarity of the cosine similarity of the feature vector is: Among them, χ represents the weighted fusion similarity; λ i , j , k represents the eigenvalue, where λ i , j , k denote the i-th eigenvalue, j-th eigenvalue, and k-th eigenvalue respectively; CS(d i , d j ) represents the cosine similarity of feature vectors.

Images