Method, device, equipment and storage medium for determining DC bias information of transformer

Abstract

This application relates to a method, apparatus, device, and storage medium for determining DC bias magnetic information of a transformer. The method comprises: obtaining a vibration signal sequence of the transformer; extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain characteristics, frequency domain characteristics, or time-frequency domain characteristics; and determining the DC bias magnetic information of the transformer based on the characteristic parameters. This method can accurately determine the DC bias magnetic information of the transformer.

Description

Technical Field

[0001] The present application relates to the field of signal processing technology, and in particular to a method, device, equipment and storage medium for determining DC bias magnetic information of a transformer. Background Art

[0002] As the scale of power grids continues to expand, abnormal phenomena such as increased vibration and noise caused by DC bias in transformers are appearing in more and more areas. DC bias in transformers can have serious impacts on both the transformer itself and the power system.

[0003] Currently, the DC bias state of a transformer is primarily assessed by calculating the excitation current through simulation models. However, this current is difficult to measure during actual transformer operation and cannot accurately monitor the DC bias state of the transformer. Therefore, accurately determining the DC bias information of the transformer is crucial. Summary of the Invention

[0004] Based on this, it is necessary to provide a method, device, equipment and storage medium for determining the DC bias magnetic information of a transformer that can accurately determine the DC bias magnetic information of the transformer in order to address the above technical problems.

[0005] In a first aspect, the present application provides a method for determining DC bias magnetic information of a transformer, the method comprising:

[0006] Obtaining the vibration signal sequence of the transformer;

[0007] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0008] The DC bias information of the transformer is determined according to the characteristic parameters.

[0009] In one embodiment, determining the DC bias information of the transformer according to the characteristic parameters includes:

[0010] Inputting the time domain features in the characteristic parameters into a first neural network to obtain a first probability distribution of the DC bias level of the transformer;

[0011] Inputting the frequency domain features in the characteristic parameters into a second neural network to obtain a second probability distribution of the DC bias level of the transformer;

[0012] Inputting the time-frequency domain features in the characteristic parameters into a third neural network to obtain a third probability distribution of the DC bias level of the transformer;

[0013] The DC bias level of the transformer is determined according to the first probability distribution, the second probability distribution, and the third probability distribution.

[0014] In one embodiment, determining the DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution includes:

[0015] According to predetermined weight values, the first probability distribution, the second probability distribution, and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include the weight values ​​of the time domain features, the weight values ​​of the frequency domain features, and the weight values ​​of the time-frequency domain features;

[0016] The DC bias level of the transformer is determined according to the fused probability distribution.

[0017] In one embodiment, the method further comprises:

[0018] According to the relative importance of the time domain features, the frequency domain features and the time-frequency domain features, a hierarchical analysis method is used to determine the weight value of the time domain features, the weight value of the frequency domain features and the weight value of the time-frequency domain features.

[0019] In one embodiment, the method further comprises:

[0020] determining an early warning mode according to the DC bias magnetic level of the transformer;

[0021] According to the warning method, output warning prompt information.

[0022] In one embodiment, extracting characteristic parameters of the vibration signal sequence includes at least two of the following:

[0023] performing a vibration amplitude operation on the vibration signal sequence, and determining a time domain feature in a characteristic parameter of the vibration signal sequence based on the amplitude operation result;

[0024] Performing a frequency domain transformation on the vibration signal sequence, and determining the frequency domain features in the characteristic parameters according to the frequency domain transformation result;

[0025] The vibration signal sequence is subjected to wavelet transformation, and the time-frequency domain features of the characteristic parameters are determined according to the wavelet transformation result.

[0026] In a second aspect, the present application further provides a device for determining DC bias magnetic information of a transformer, the device comprising:

[0027] A signal acquisition module, used to acquire a vibration signal sequence of the transformer;

[0028] A parameter extraction module, configured to extract characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0029] An information determination module is used to determine the DC bias information of the transformer according to the characteristic parameters.

[0030] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0031] Obtaining the vibration signal sequence of the transformer;

[0032] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0033] The DC bias information of the transformer is determined according to the characteristic parameters.

[0034] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0035] Obtaining the vibration signal sequence of the transformer;

[0036] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0037] The DC bias information of the transformer is determined according to the characteristic parameters.

[0038] In a fifth aspect, the present application further provides a computer program product, the computer program product comprising a computer program, which, when executed by a processor, implements the following steps:

[0039] Obtaining the vibration signal sequence of the transformer;

[0040] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0041] The DC bias information of the transformer is determined according to the characteristic parameters.

[0042] The above-mentioned method, device, computer equipment, storage medium and computer program product for determining the DC bias magnetic information of the transformer can accurately determine the DC bias magnetic information of the transformer by extracting at least two characteristic parameters from the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of the vibration signal sequence of the transformer, and based on the extracted characteristic parameters. In the above scheme, on the one hand, the present application introduces a vibration signal sequence to determine the DC bias magnetic information of the transformer. Since the vibration signal sequence of the transformer can be obtained conveniently and in real time, the effect of determining the DC bias magnetic information of the transformer in real time is achieved; on the other hand, by combining at least two characteristic parameters from the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics to evaluate the DC bias magnetic information of the transformer, the determination of the DC bias magnetic information of the transformer can be made more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 FIG. 1 is an application environment diagram of a method for determining DC bias magnetic information of a transformer in one embodiment;

[0044] Figure 2 1 is a flow chart of a method for determining DC bias magnetic information of a transformer in one embodiment;

[0045] Figure 3 A schematic diagram of a flow chart for determining DC bias magnetic information of a transformer in another embodiment;

[0046] Figure 4 FIG1 is a schematic diagram showing the principle of determining DC bias magnetic information of a transformer in one embodiment;

[0047] Figure 5 is a structural block diagram of a device for determining DC bias magnetic information of a transformer in one embodiment;

[0048] Figure 6 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0050] The method for determining DC bias information of a transformer provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the terminal 102 communicates with the server 104 via a network. The data storage system can store data that the server 104 needs to process. The data storage system can be integrated on the server 104 or placed on the cloud or other network servers.

[0051] Optionally, the method of this embodiment can be applied to Figure 1 The server shown can also be applied to Figure 1 The terminal with relatively powerful computing power shown in the figure can also be applied to a system including a terminal and a server, and implemented through the interaction between the terminal and the server. For example, a user can send a DC bias information determination request to the server through the terminal held by the user. When the server obtains the DC bias information determination request sent by the terminal, it obtains the vibration signal sequence of the transformer and extracts at least two characteristic parameters of the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of the vibration signal sequence, and then determines the DC bias information of the transformer based on the extracted characteristic parameters. After that, the server can send the determined DC bias information to the terminal for the user to view, etc.

[0052] Terminal 102 may include, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, portable wearable devices, and dedicated simulation devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, smart car devices, etc. Portable wearable devices may include smart watches, smart bracelets, head-mounted devices, etc. Server 104 may be implemented as a standalone server or a server cluster consisting of multiple servers.

[0053] In one embodiment, Figure 2 As shown, a method for determining DC bias information of a transformer is provided, and the method is applied to Figure 1 Specifically, the method may include the following steps:

[0054] S201: Acquire a vibration signal sequence of a transformer.

[0055] The transformer in S201 can be any transformer for which DC bias magnetic information needs to be determined. Optionally, in this embodiment, a vibration sensor can be installed on the transformer housing to collect the transformer's vibration signal sequence. Furthermore, the vibration sensor in this embodiment can be a velocity-type vibration sensor, such as a velocity or acceleration vibration sensor.

[0056] Specifically, when it is determined that the DC bias magnetic information of the transformer needs to be evaluated, a vibration sensor installed in the transformer box can be used to collect the vibration signal sequence of the transformer at a set sampling frequency (such as 5 kHz).

[0057] The need to evaluate the DC bias information of the transformer may be determined by detecting that the current time meets a set evaluation period, or by obtaining a request for determining the DC bias information of the transformer, etc. This embodiment does not limit this.

[0058] S202: Extract characteristic parameters of the vibration signal sequence.

[0059] Optionally, the characteristic parameters in this embodiment include at least two of time domain characteristics, frequency domain characteristics, or time-frequency domain characteristics.

[0060] It should be noted that the vibration signal sequence collected in this embodiment is a discrete time domain signal, so the vibration signal sequence can be directly subjected to time domain analysis to extract the time domain features of the vibration signal sequence.

[0061] Furthermore, the vibration signal sequence can be subjected to frequency domain analysis to extract the frequency domain features of the vibration signal sequence; or the vibration signal sequence can be subjected to time-frequency domain analysis (for example, the vibration signal sequence can be subjected to wavelet transform and the wavelet transform results can be analyzed) to extract the time-frequency domain features of the vibration signal sequence.

[0062] In one embodiment, feature parameters of the vibration signal sequence may be extracted based on machine learning. For example, the collected vibration signal sequence may be input into a pre-trained feature extraction network, which then outputs the feature parameters of the vibration signal sequence.

[0063] S203: Determine the DC bias information of the transformer according to the characteristic parameters.

[0064] In this embodiment, the DC bias information of the transformer may include the DC bias level of the transformer. Optionally, the DC bias level of the transformer may be any one of the four levels A, B, C, and D, which are set in advance by statistically analyzing the severity of the DC bias of a large number of transformers. Among them, level A indicates that there is no DC bias in the transformer; level B indicates that there is a DC bias in the transformer, but the DC bias has little effect on the transformer, and attention should be paid to the DC bias of the transformer; level C indicates that the DC bias has a greater impact on the normal operation of the transformer, and the monitoring of the DC bias of the transformer should be strengthened; level D indicates that the DC bias has seriously affected the normal operation of the transformer, and the DC bias should be suppressed in a timely manner.

[0065] Optionally, after extracting the characteristic parameters of the vibration signal sequence, this embodiment can input the extracted characteristic parameters into a pre-trained neural network model, and the model outputs the DC bias level of the transformer.

[0066] Furthermore, after determining the DC bias level of the transformer, it is possible to quickly determine whether the transformer has a DC bias phenomenon and the extent of the impact of the DC bias phenomenon on the transformer based on the DC bias level of the transformer, thereby facilitating the rapid customization of a corresponding solution strategy for the DC bias phenomenon.

[0067] The above-mentioned method for determining the DC bias magnetic information of the transformer can accurately determine the DC bias magnetic information of the transformer by extracting at least two characteristic parameters from the time domain characteristics, frequency domain characteristics, and time-frequency domain characteristics of the vibration signal sequence of the transformer, and based on the extracted characteristic parameters. In the above scheme, on the one hand, the present application introduces a vibration signal sequence to determine the DC bias magnetic information of the transformer. Since the vibration signal sequence of the transformer can be obtained conveniently and in real time, the effect of determining the DC bias magnetic information of the transformer in real time is achieved; on the other hand, by combining at least two characteristic parameters from the time domain characteristics, frequency domain characteristics, and time-frequency domain characteristics to evaluate the DC bias magnetic information of the transformer, the determination of the DC bias magnetic information of the transformer can be made more accurate.

[0068] For example, based on the above embodiment, the process of S202, i.e., extracting characteristic parameters of the vibration signal sequence, is further explained in detail. Specifically, the characteristic parameters of the vibration signal sequence extracted may include at least two of the following:

[0069] The first item is to perform a vibration amplitude operation on the vibration signal sequence, and determine the time domain characteristics of the characteristic parameters of the vibration signal sequence based on the amplitude operation result.

[0070] Specifically, this embodiment may calculate the mean and maximum value of the vibration amplitude of the vibration signal sequence; and use the mean and / or maximum value of the vibration amplitude as the time domain feature in the feature parameters of the vibration signal sequence.

[0071] The second item is to perform frequency domain transformation on the vibration signal sequence and determine the frequency domain features in the characteristic parameters based on the frequency domain transformation results.

[0072] Specifically, this embodiment can perform Fourier transform on the vibration signal sequence to obtain the spectrum of the vibration signal sequence; extract the fundamental wave amplitude and the amplitudes of each harmonic from the spectrum; determine the spectrum complexity of the vibration signal sequence and the parity ratio of the vibration signal sequence based on the fundamental wave amplitude and the amplitudes of each harmonic; and use the spectrum complexity and / or parity ratio as frequency domain features in the characteristic parameters of the vibration signal sequence.

[0073] For example, since the operating voltage of a transformer is typically 50 Hz, the amplitude of the 50 nHz frequency component can be extracted from the spectrum of the vibration signal sequence. Here, n is a value of 1, 2, …, 20. The amplitude of the frequency component when n is 1 (i.e., 50 Hz) is the fundamental amplitude, also known as the first harmonic amplitude; the amplitude of the frequency component when n is 2 (i.e., 100 Hz) is the second harmonic amplitude; …, the amplitude of the frequency component when n is 20 (i.e., 1 kHz) is the twentieth harmonic amplitude.

[0074] Furthermore, the total amplitude of the odd-frequency components and the total amplitude of the even-frequency components in the spectrum of the vibration signal sequence can be determined based on the amplitude of the fundamental wave and the amplitude of each harmonic, and the ratio of the two total amplitudes can be used as the odd-even ratio of the vibration signal sequence. The odd-frequency components in the spectrum of the vibration signal sequence are the odd-frequency components of 50Hz, i.e., 50Hz, 150Hz... and 950Hz; correspondingly, the even-frequency components in the spectrum of the vibration signal sequence are the even-frequency components of 50Hz, i.e., 100Hz, 200Hz... and 1kHz.

[0075] That is, the odd-even ratio of the vibration signal sequence can be determined by the following formula 1, where H 2u is the amplitude of the even-harmonic frequency component of 50 Hz; H 2u-1 is the amplitude of the odd-order frequency component of 50 Hz.

[0076]

[0077] Furthermore, the total energy of the vibration signal sequence can be determined based on the fundamental wave amplitude and the amplitudes of each harmonic, and the proportion of each harmonic amplitude in the total energy can be calculated; the spectral complexity of the vibration signal sequence can be determined based on the proportion of each harmonic amplitude in the total energy.

[0078] That is, the spectrum complexity F of the vibration signal sequence can be determined by the following formula 2 and formula 3, where R v is the proportion of the 50Hz vth harmonic amplitude in the vibration signal sequence to the total energy (i.e. ) proportion.

[0079]

[0080]

[0081] The third item is to perform wavelet transform on the vibration signal sequence and determine the time-frequency domain features in the characteristic parameters based on the wavelet transform results.

[0082] Specifically, this embodiment can use different transformation scales to perform wavelet transform on the vibration signal sequence to obtain the wavelet energy of the vibration signal sequence at different transformation scales. For example, the wavelet energy E of the vibration signal sequence at the kth transformation scale is k It can be determined by the following formula 4, where k can be 1, 2, 3...M (M is a positive integer), N is the length of the vibration signal sequence, and d lk is the lth wavelet coefficient of the vibration signal sequence at the kth transformation scale.

[0083]

[0084] Afterwards, the wavelet energy entropy of the vibration signal sequence can be determined based on the wavelet energy of the vibration signal sequence at different transformation scales. For example, the sum of the wavelet energy E at different transformation scales can be calculated, and the energy ratio p at different transformation scales can be calculated using the following formula 5: k , and then according to the energy ratio under different transformation scales, the following formula 6 is used to calculate the wavelet energy entropy W of the vibration signal sequence EE .in,

[0085] p k =E k / E (Formula 5)

[0086]

[0087] Furthermore, by using different transformation scales to perform wavelet transform on the vibration signal sequence, wavelet coefficients at different transformation scales can be obtained. For example, using M transformation scales to perform wavelet transform on the vibration signal sequence, for each transformation scale, wavelet coefficients of length N can be obtained, and then the wavelet coefficients of M transformation scales can form an M×N matrix D M×N . For the matrix D M×N By performing singular value decomposition, the wavelet singular entropy of the vibration signal sequence can be obtained.

[0088] Afterwards, the wavelet energy entropy and / or wavelet singular entropy of the vibration signal sequence may be used as time-frequency domain features in the feature parameters.

[0089] It can be understood that, in this embodiment, the above two or more items can be used to determine the characteristic parameters of the transformer, thereby improving the flexibility of determining the characteristic parameters.

[0090] Figure 3 FIG. 1 is a flow chart of determining DC bias information of a transformer in another embodiment. Figure 4 This is a schematic diagram of the principle of determining the DC bias magnetic information of a transformer in one embodiment. Based on the above embodiment, this embodiment provides an optional method for determining the DC bias magnetic information of a transformer based on a neural network in a scenario where the characteristic parameters include time domain characteristics, frequency domain characteristics and time-frequency domain characteristics. Figure 3 and Figure 4 As shown, the following steps may be specifically included:

[0091] S301: Acquire a vibration signal sequence of a transformer.

[0092] S302: Extract characteristic parameters of the vibration signal sequence.

[0093] S303: Input the time domain features in the feature parameters into a first neural network to obtain a first probability distribution of the DC bias level of the transformer.

[0094] S304: Input the frequency domain features in the feature parameters into a second neural network to obtain a second probability distribution of the DC bias level of the transformer.

[0095] S305 , inputting the time-frequency domain features in the characteristic parameters into a third neural network to obtain a third probability distribution of the DC bias level of the transformer.

[0096] S306 : Determine a DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution.

[0097] Optionally, to reduce the complexity of training, the first neural network, the second neural network, and the third neural network in this embodiment can be pre-trained using training samples. For example, the sample time domain features can be input into the first initial network, the sample frequency domain features can be input into the second initial network, and the sample time and frequency domain features can be input into the third initial network. Thereafter, a training loss is determined based on the first predicted probability distribution output by the first initial network, the second predicted probability distribution output by the second initial network, the third predicted probability distribution output by the third initial network, and the sample DC bias level of the sample vibration signal sequence. The first initial network, the second initial network, and the third initial network are jointly trained using the training loss to obtain the trained first neural network, the second neural network, and the third neural network.

[0098] To further reduce computational complexity, in this embodiment, the sample time domain features may be normalized before being input into the first initial network; similarly, the sample frequency domain features may be normalized before being input into the second initial network; and the sample time and frequency domain features may be normalized before being input into the third initial network.

[0099] Optional, such as Figure 4 As shown, after determining the time domain characteristics, frequency domain characteristics, and time-frequency domain characteristics of the transformer vibration signal sequence, the time domain characteristics of the vibration signal sequence (such as the maximum and mean values ​​of the vibration amplitude) can be input into the first neural network to obtain a first probability distribution; at the same time, the frequency domain characteristics of the vibration signal sequence (such as spectral complexity and odd-even ratio) can be input into the second neural network to obtain a second probability distribution; and the time-frequency domain characteristics of the vibration signal sequence (such as wavelet energy entropy and wavelet singular entropy) can be input into the third neural network to obtain a third probability distribution. Among them, the first probability distribution is the distribution of the predicted DC bias magnetic levels of the transformer, which are A, B, C, and D respectively; similarly, the second probability distribution and the third probability distribution are also the distribution of the predicted DC bias magnetic levels of the transformer, which are A, B, C, and D respectively.

[0100] Afterwards, the first probability distribution, the second probability distribution, and the third probability distribution can be fused. For example, the first probability distribution, the second probability part, and the third probability distribution can be added together to obtain the fused probability distribution of the transformer, and then the DC bias level of the transformer can be determined based on the fused probability part of the transformer. For example, the final probability distribution of the transformer is [0, 0.7, 0.1, 0.2]; wherein 0 represents the probability distribution of the predicted DC bias level of the transformer being A, 0.7 represents the probability distribution of the predicted DC bias level of the transformer being B, 0.1 represents the probability distribution of the predicted DC bias level of the transformer being C, and 0.2 represents the probability distribution of the predicted DC bias level of the transformer being D. At this point, the DC bias level corresponding to the maximum value in the final probability distribution can be directly used as the DC bias level of the transformer. That is, the level B corresponding to 0.7 is used as the DC bias level of the transformer.

[0101] The technical solution provided in this embodiment uses a neural network to process the time, frequency, and time-frequency domain characteristics of the vibration signal sequence to evaluate the transformer's DC bias information, further improving the accuracy of determining the transformer's DC bias information. Furthermore, the introduction of DC bias levels in this embodiment further facilitates the development of subsequent DC bias suppression strategies.

[0102] To further improve the accuracy of determining DC bias information, S306 is further explained in detail based on the above embodiment. Optionally, S306 can be implemented as follows: based on predetermined weight values, the first probability distribution, the second probability distribution, and the third probability distribution are fused to obtain a fused probability distribution; and the DC bias level of the transformer is determined based on the fused probability distribution.

[0103] The predetermined weight values ​​include the weight values ​​of time domain features, the weight values ​​of frequency domain features, and the weight values ​​of time-frequency domain features.

[0104] Specifically, the weighted value of the time-domain feature can be multiplied by the first probability distribution to obtain a first product; the weighted value of the frequency-domain feature can be multiplied by the second probability distribution to obtain a second product; and the weighted value of the time-frequency domain feature can be multiplied by the third probability distribution to obtain a third product. The sum of the first, second, and third products is then used as a fused probability distribution. The DC bias level of the transformer can then be determined based on the fused probability distribution.

[0105] It can be understood that, by introducing the weight value in this embodiment, the fusion between the first probability distribution, the second probability distribution and the third probability distribution can be made more reasonable, thereby making the determination of the DC bias magnetic level more accurate.

[0106] Exemplarily, to ensure the rationality of weight value allocation, based on the above embodiment, an optional method for determining the weight values ​​of time domain features, the weight values ​​of frequency domain features, and the weight values ​​of time-frequency domain features is provided. Specifically, the weight values ​​of time domain features, the weight values ​​of frequency domain features, and the weight values ​​of time-frequency domain features can be determined using a hierarchical analysis method based on the relative importance of time domain features, frequency domain features, and the relative importance of each other among time-frequency domain features.

[0107] The relative importance of the time domain features, the frequency domain features, and the time-frequency domain features may be determined by subjective judgment or based on a certain judgment strategy, which is not limited in this embodiment.

[0108] Optionally, a comparison matrix can be established based on the relative importance of time domain features, frequency domain features, and time-frequency domain features. ij =0 means that the factor corresponding to row i is less important than the factor corresponding to column j; ij =1 means that the factor corresponding to the i-th row and the factor corresponding to the j-th column are equally important; a ij =2 means that the factor corresponding to row i is more important than the factor corresponding to column j. Furthermore, the importance of each factor relative to itself is 1.

[0109] Assume that there are three factors, namely time domain features, frequency domain features and time-frequency domain features. At this time, the comparison matrix A has 3 rows, the first row corresponds to the factor of time domain features, the second row corresponds to the factor of frequency domain features, and the third row corresponds to the factor of time-frequency domain features; the comparison matrix A also has 3 columns, the first column corresponds to the factor of time domain features, the second column corresponds to the factor of frequency domain features, and the third column corresponds to the factor of time-frequency domain features.

[0110] Then, based on the relative importance of time domain features, frequency domain features, and each of the time-frequency domain features, the following comparison matrix A1 can be established:

[0111]

[0112] Based on the comparison matrix, determine the importance ranking index of each feature among the time domain features, frequency domain features, and time-frequency domain features. For example, the importance ranking index of the time domain features, the importance ranking index of the frequency domain features, and the importance ranking index of the time-frequency domain features can be determined using the following formula 7. Wherein, L represents the number of columns in the comparison matrix A. For example, L is 3 in A1.

[0113]

[0114] For example, the comparison matrix is ​​A1, and using Formula 7, we can obtain the importance ranking index of time domain features r1=3, the importance ranking index of frequency domain features r2=1, and the importance ranking index of time-frequency domain features r3=5.

[0115] Afterwards, the judgment matrix B is constructed based on the importance ranking index of the time-frequency feature, the importance ranking index of the frequency domain feature, and the importance ranking index of the time-frequency domain feature. Among them, the element b in the judgment matrix B is ij It can be determined by the following formula 8. Where, k m =r max / r min , r max is the maximum value in the importance ranking index, r min is the minimum value in the importance ranking index.

[0116]

[0117] For example, if the importance ranking index of time domain features r1=3, the importance ranking index of frequency domain features r2=1, and the importance ranking index of time-frequency domain features r3=5 are substituted into the above formula 8, the following judgment matrix B1 can be obtained:

[0118]

[0119] After obtaining the judgment matrix B, the element b in the judgment matrix B is ij The optimal transfer matrix C of the judgment matrix B is obtained by processing. Specifically, the optimal transfer matrix C can be determined using the following formula 9.

[0120]

[0121] For example, based on the above judgment B1, the following optimal transfer matrix C1 can be obtained:

[0122]

[0123] Next, the optimal transfer matrix C can be processed using the following formula 10 to obtain the quasi-optimal consistent matrix T.

[0124]

[0125] For example, based on the optimal transfer matrix C1, the following quasi-optimal consistency matrix T1 can be obtained:

[0126]

[0127] After that, the quasi-optimal consistent matrix is ​​normalized to obtain the weight value of the time domain feature, the weight value of the frequency domain feature, and the weight value of the time-frequency domain feature. iIt can be determined by the following formula 11, where i can take values ​​of 1, 2, ...x (x is a positive integer, that is, the number of factors), such as 3.

[0128]

[0129] It can be understood that in this embodiment, the improved analytic hierarchy process is used to distribute weights among time domain features, frequency domain features, and time-frequency domain features, so that the DC bias magnetic level can be determined more accurately.

[0130] For example, based on the above embodiments, as an implementable embodiment of the present application, after determining the DC bias level of the transformer, the warning method can also be determined according to the DC bias level of the transformer; and according to the warning method, the warning prompt information can be output.

[0131] Optionally, different DC bias levels can be pre-set, corresponding to different warning methods. For example, for level A, the warning method may be a text message prompt; for level B, the warning method may be a text message and voice prompt; for level C, the warning method may be a voice and vibration prompt; for level D, the warning method may be a continuous ringing and vibration prompt.

[0132] Specifically, the warning mode can be determined according to the predetermined correspondence between the DC bias level and the warning mode and the DC bias level of the transformer; and then, the warning prompt information is output according to the determined warning mode.

[0133] Furthermore, the content of the warning message may vary depending on the warning method. For example, for a text message, the warning message could be "The transformer is operating normally"; for a text message and voice prompt, the warning message could be "Pay attention to DC bias in the transformer"; for a voice and vibration prompt, the warning message could be "Reinforce monitoring of DC bias in the transformer"; for a continuous ringing and vibration warning, the warning message could be "Suppress DC bias in the transformer promptly"; and so on.

[0134] It is understandable that this embodiment uses different warning modes and warning prompt information to provide a warning of the DC bias level, which can facilitate intuitive determination of the severity of the transformer DC bias.

[0135] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0136] Based on the same inventive concept, embodiments of the present application further provide a device for determining DC bias magnetic information of a transformer, for implementing the aforementioned method for determining DC bias magnetic information of a transformer. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of the embodiments of one or more devices for determining DC bias magnetic information of a transformer provided below can be found in the aforementioned limitations of the method for determining DC bias magnetic information of a transformer, and are not further elaborated here.

[0137] In one embodiment, Figure 5 As shown, a device 500 for determining DC bias magnetic information of a transformer is provided, comprising: a signal acquisition module 510, a parameter extraction module 520 and an information determination module 530, wherein:

[0138] A signal acquisition module 510 is used to acquire a vibration signal sequence of the transformer;

[0139] The parameter extraction module 520 is used to extract characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0140] The information determination module 530 is used to determine the DC bias information of the transformer according to the characteristic parameters.

[0141] In one embodiment, the information determination module 530 includes:

[0142] A first determining unit is configured to input the time domain features in the feature parameters into a first neural network to obtain a first probability distribution of the DC bias level of the transformer;

[0143] A second determining unit is configured to input the frequency domain features in the characteristic parameters into a second neural network to obtain a second probability distribution of the DC bias level of the transformer;

[0144] a third determining unit, configured to input the time-frequency domain features in the characteristic parameters into a third neural network to obtain a third probability distribution of the DC bias level of the transformer;

[0145] The fourth determining unit is configured to determine a DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution.

[0146] In one embodiment, the fourth determining unit is specifically configured to:

[0147] According to predetermined weight values, the first probability distribution, the second probability distribution, and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include weight values ​​of time domain features, weight values ​​of frequency domain features, and weight values ​​of time-frequency domain features;

[0148] According to the fusion probability distribution, the DC bias level of the transformer is determined.

[0149] Exemplarily, the above device further includes:

[0150] The weight determination module is used to determine the weight value of the time domain feature, the weight value of the frequency domain feature and the weight value of the time-frequency domain feature according to the relative importance of the time domain feature, the frequency domain feature and the relative importance of each pair of features in the time-frequency domain feature, using the hierarchical analysis method.

[0151] Exemplarily, the above device further includes:

[0152] A mode determination module is used to determine the warning mode according to the DC bias magnetic level of the transformer;

[0153] The information output module is used to output warning prompt information according to the warning method.

[0154] In one embodiment, the parameter extraction module 520 may specifically perform at least two of the following:

[0155] Performing a vibration amplitude operation on the vibration signal sequence, and determining a time domain feature in a characteristic parameter of the vibration signal sequence based on the amplitude operation result;

[0156] Performing frequency domain transformation on the vibration signal sequence, and determining the frequency domain features in the characteristic parameters according to the frequency domain transformation results;

[0157] The vibration signal sequence is subjected to wavelet transform, and the time-frequency domain features in the characteristic parameters are determined according to the wavelet transform results.

[0158] Each module in the device for determining DC bias magnetic information of a transformer can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each module.

[0159] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for determining DC bias magnetic information of a transformer is implemented.

[0160] Those skilled in the art will understand that Figure 6 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 shown in the figure, or combine certain components, or have a different component arrangement.

[0161] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0162] Obtaining the vibration signal sequence of the transformer;

[0163] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0164] According to the characteristic parameters, the DC bias information of the transformer is determined.

[0165] In one embodiment, when the processor executes the logic of determining the DC bias information of the transformer according to the characteristic parameters in the computer program, the processor specifically implements the following steps:

[0166] The time domain features in the characteristic parameters are input into the first neural network to obtain a first probability distribution of the DC bias level of the transformer; the frequency domain features in the characteristic parameters are input into the second neural network to obtain a second probability distribution of the DC bias level of the transformer; the time and frequency domain features in the characteristic parameters are input into the third neural network to obtain a third probability distribution of the DC bias level of the transformer; and the DC bias level of the transformer is determined based on the first probability distribution, the second probability distribution and the third probability distribution.

[0167] In one embodiment, when the processor executes the logic of determining the DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution in the computer program, the processor specifically implements the following steps:

[0168] According to predetermined weight values, the first probability distribution, the second probability distribution and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include weight values ​​of time domain features, weight values ​​of frequency domain features and weight values ​​of time-frequency domain features; according to the fused probability distribution, the DC bias level of the transformer is determined.

[0169] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0170] According to the relative importance of time domain features, frequency domain features and the relative importance of each other in time-frequency domain features, the hierarchical analysis method is used to determine the weight value of time domain features, the weight value of frequency domain features and the weight value of time-frequency domain features.

[0171] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0172] The warning mode is determined according to the DC bias magnetic level of the transformer; and the warning prompt information is output according to the warning mode.

[0173] In one embodiment, when executing the logic of extracting characteristic parameters of the vibration signal sequence in the computer program, the processor may further implement at least two of the following:

[0174] Perform vibration amplitude operation on the vibration signal sequence, and determine the time domain characteristics of the characteristic parameters of the vibration signal sequence based on the amplitude operation result; perform frequency domain transformation on the vibration signal sequence, and determine the frequency domain characteristics of the characteristic parameters based on the frequency domain transformation result; perform wavelet transformation on the vibration signal sequence, and determine the time-frequency domain characteristics of the characteristic parameters based on the wavelet transformation result.

[0175] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0176] Obtaining the vibration signal sequence of the transformer;

[0177] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0178] According to the characteristic parameters, the DC bias information of the transformer is determined.

[0179] In one embodiment, when the processor executes the logic of determining the DC bias information of the transformer according to the characteristic parameters in the computer program, the processor specifically implements the following steps:

[0180] The time domain features in the characteristic parameters are input into the first neural network to obtain a first probability distribution of the DC bias level of the transformer; the frequency domain features in the characteristic parameters are input into the second neural network to obtain a second probability distribution of the DC bias level of the transformer; the time and frequency domain features in the characteristic parameters are input into the third neural network to obtain a third probability distribution of the DC bias level of the transformer; and the DC bias level of the transformer is determined based on the first probability distribution, the second probability distribution and the third probability distribution.

[0181] In one embodiment, when the processor executes the logic of determining the DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution in the computer program, the processor specifically implements the following steps:

[0182] According to predetermined weight values, the first probability distribution, the second probability distribution and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include weight values ​​of time domain features, weight values ​​of frequency domain features and weight values ​​of time-frequency domain features; according to the fused probability distribution, the DC bias level of the transformer is determined.

[0183] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0184] According to the relative importance of time domain features, frequency domain features and the relative importance of each other in time-frequency domain features, the hierarchical analysis method is used to determine the weight value of time domain features, the weight value of frequency domain features and the weight value of time-frequency domain features.

[0185] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0186] The warning mode is determined according to the DC bias magnetic level of the transformer; and the warning prompt information is output according to the warning mode.

[0187] In one embodiment, when executing the logic of extracting characteristic parameters of the vibration signal sequence in the computer program, the processor may further implement at least two of the following:

[0188] Perform vibration amplitude operation on the vibration signal sequence, and determine the time domain characteristics of the characteristic parameters of the vibration signal sequence based on the amplitude operation result; perform frequency domain transformation on the vibration signal sequence, and determine the frequency domain characteristics of the characteristic parameters based on the frequency domain transformation result; perform wavelet transformation on the vibration signal sequence, and determine the time-frequency domain characteristics of the characteristic parameters based on the wavelet transformation result.

[0189] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0190] Obtaining the vibration signal sequence of the transformer;

[0191] Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include at least two of time domain features, frequency domain features, or time-frequency domain features;

[0192] According to the characteristic parameters, the DC bias information of the transformer is determined.

[0193] In one embodiment, when the processor executes the logic of determining the DC bias information of the transformer according to the characteristic parameters in the computer program, the processor specifically implements the following steps:

[0194] The time domain features in the characteristic parameters are input into the first neural network to obtain a first probability distribution of the DC bias level of the transformer; the frequency domain features in the characteristic parameters are input into the second neural network to obtain a second probability distribution of the DC bias level of the transformer; the time and frequency domain features in the characteristic parameters are input into the third neural network to obtain a third probability distribution of the DC bias level of the transformer; and the DC bias level of the transformer is determined based on the first probability distribution, the second probability distribution and the third probability distribution.

[0195] In one embodiment, when the processor executes the logic of determining the DC bias level of the transformer according to the first probability distribution, the second probability distribution, and the third probability distribution in the computer program, the processor specifically implements the following steps:

[0196] According to predetermined weight values, the first probability distribution, the second probability distribution and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include weight values ​​of time domain features, weight values ​​of frequency domain features and weight values ​​of time-frequency domain features; according to the fused probability distribution, the DC bias level of the transformer is determined.

[0197] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0198] According to the relative importance of time domain features, frequency domain features and the relative importance of each other in time-frequency domain features, the hierarchical analysis method is used to determine the weight value of time domain features, the weight value of frequency domain features and the weight value of time-frequency domain features.

[0199] In one embodiment, when executing the computer program, the processor may further implement the following steps:

[0200] The warning mode is determined according to the DC bias magnetic level of the transformer; and the warning prompt information is output according to the warning mode.

[0201] In one embodiment, when executing the logic of extracting characteristic parameters of the vibration signal sequence in the computer program, the processor may further implement at least two of the following:

[0202] Perform vibration amplitude operation on the vibration signal sequence, and determine the time domain characteristics of the characteristic parameters of the vibration signal sequence based on the amplitude operation result; perform frequency domain transformation on the vibration signal sequence, and determine the frequency domain characteristics of the characteristic parameters based on the frequency domain transformation result; perform wavelet transformation on the vibration signal sequence, and determine the time-frequency domain characteristics of the characteristic parameters based on the wavelet transformation result.

[0203] It should be noted that the vibration signal sequence involved in this application is obtained after full authorization from all parties.

[0204] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0205] The technical features of the above embodiments can be combined arbitrarily. 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.

[0206] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for determining DC bias magnetic information of a transformer, characterized in that: The method comprises: Obtaining the vibration signal sequence of the transformer; Extracting characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include time domain characteristics, frequency domain characteristics, and time-frequency domain characteristics; the time domain characteristics include the maximum value and / or mean value of the vibration amplitude; the frequency domain characteristics include spectrum complexity and / or odd-even ratio; and the time-frequency domain characteristics include wavelet energy entropy and / or wavelet singular entropy; Inputting the time domain features in the characteristic parameters into a first neural network to obtain a first probability distribution of the DC bias level of the transformer; Inputting the frequency domain features in the characteristic parameters into a second neural network to obtain a second probability distribution of the DC bias level of the transformer; Inputting the time-frequency domain features in the characteristic parameters into a third neural network to obtain a third probability distribution of the DC bias level of the transformer; According to predetermined weight values, the first probability distribution, the second probability distribution, and the third probability distribution are fused to obtain a fused probability distribution; the predetermined weight values ​​include the weight values ​​of the time domain features, the weight values ​​of the frequency domain features, and the weight values ​​of the time-frequency domain features; determining a DC bias level of the transformer according to the fused probability distribution; The probability distribution is the distribution of the predicted DC bias magnetic levels of the transformer, which are A, B, C, and D. The four levels A, B, C, and D are used to represent the severity of the DC bias magnetic level of the transformer. The weight value of the time domain feature, the weight value of the frequency domain feature and the weight value of the time-frequency domain feature are determined in the following manner: based on the relative importance of each of the time domain feature, the frequency domain feature and the time-frequency domain feature, the hierarchical analysis method is used to determine the weight value of the time domain feature, the weight value of the frequency domain feature and the weight value of the time-frequency domain feature.

2. The method according to claim 1, characterized in that The method further comprises: determining an early warning mode according to the DC bias magnetic level of the transformer; According to the warning method, output warning prompt information.

3. The method according to claim 1, characterized in that The step of extracting characteristic parameters of the vibration signal sequence includes at least two of the following: performing a vibration amplitude operation on the vibration signal sequence, and determining a time domain feature in a characteristic parameter of the vibration signal sequence based on the amplitude operation result; Performing a frequency domain transformation on the vibration signal sequence, and determining the frequency domain features in the characteristic parameters according to the frequency domain transformation result; The vibration signal sequence is subjected to wavelet transformation, and the time-frequency domain features of the characteristic parameters are determined according to the wavelet transformation result.

4. A device for determining DC bias magnetic information of a transformer, characterized in that: The device comprises: A signal acquisition module, used to acquire a vibration signal sequence of the transformer; a parameter extraction module, configured to extract characteristic parameters of the vibration signal sequence; wherein the characteristic parameters include time domain characteristics, frequency domain characteristics, and time-frequency domain characteristics; the time domain characteristics include the maximum value and / or mean value of the vibration amplitude; the frequency domain characteristics include spectral complexity and / or odd-even ratio; and the time-frequency domain characteristics include wavelet energy entropy and / or wavelet singular entropy; An information determination module is configured to input the time domain features in the characteristic parameters into a first neural network to obtain a first probability distribution of the DC bias level of the transformer; input the frequency domain features in the characteristic parameters into a second neural network to obtain a second probability distribution of the DC bias level of the transformer; input the time-frequency domain features in the characteristic parameters into a third neural network to obtain a third probability distribution of the DC bias level of the transformer; fuse the first probability distribution, the second probability distribution, and the third probability distribution according to predetermined weight values ​​to obtain a fused probability distribution; and determine the DC bias level of the transformer according to the fused probability distribution; the predetermined weight values ​​include the weight values ​​of the time domain features, the weight values ​​of the frequency domain features, and the weight values ​​of the time-frequency domain features; The probability distribution is the distribution of the predicted DC bias magnetic levels of the transformer, which are A, B, C, and D. The four levels A, B, C, and D are used to represent the severity of the DC bias magnetic level of the transformer. The weight value of the time domain feature, the weight value of the frequency domain feature and the weight value of the time-frequency domain feature are determined in the following manner: based on the relative importance of each of the time domain feature, the frequency domain feature and the time-frequency domain feature, the hierarchical analysis method is used to determine the weight value of the time domain feature, the weight value of the frequency domain feature and the weight value of the time-frequency domain feature.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

7. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images