Excitation inrush current detection method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a magnetizing inrush current detection method and device, computer equipment, a readable storage medium and a program product. The method comprises the following steps: collecting a current signal of a transformer port; determining a time domain characteristic parameter, a frequency domain characteristic parameter and a time-frequency domain characteristic parameter of the current signal; determining importance parameters and correlation parameters of the time domain characteristic parameters, the frequency domain characteristic parameters and the time-frequency domain characteristic parameters; selecting a target characteristic parameter from the time domain characteristic parameter, the frequency domain characteristic parameter and the time-frequency domain characteristic parameter according to the importance parameter and the correlation parameter, and constructing a target characteristic vector of the current signal by using the target characteristic parameter; and adopting a pre-trained classification model to generate an excitation surge current detection result of the transformer port based on a comparison result of the target feature vector and a pre-stored excitation surge current feature. By adopting the method, the multi-dimensional characteristics of the current signal in the time domain, the frequency domain and the time-frequency domain can be fully utilized to improve the recognition precision and speed of the excitation surge current.

Description

Technical Field

[0001] The present application relates to the technical field of power system protection and fault detection, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for detecting an excitation inrush current. Background Art

[0002] During normal transformer operation, large inrush currents are often generated. These currents exhibit high peak values, severe waveform distortion, and contain a large number of low-order harmonic components. Because these characteristics of inrush currents are highly similar to those of short-circuit fault currents in both the time and frequency domains, traditional inrush current protection algorithms based on current amplitude, direction, or simple harmonic analysis struggle to accurately distinguish between the two. Existing inrush current identification methods often exhibit false or failed operation, particularly under conditions of frequent system faults or severe load fluctuations, posing a serious threat to the stable operation of the system and the safety and reliability of the equipment.

[0003] While some methods have attempted to improve upon this by incorporating second harmonic content criteria, wavelet transform analysis, and neural network recognition, most still suffer from reliance on a single feature, insufficient adaptability to diverse operating conditions, and poor real-time performance. This is particularly true given the significant variations in the amplitude, duration, and frequency content of magnetizing inrush current across different transformer types (e.g., oil-immersed and dry-type) and under varying excitation conditions. Traditional feature extraction and classification methods struggle to achieve both accuracy and versatility.

[0004] Therefore, there is an urgent need for a detection method that can accurately identify excitation inrush current under various working conditions. Summary of the Invention

[0005] Based on this, it is necessary to provide an excitation surge current detection method, device, computer equipment, computer-readable storage medium and computer program product that can quickly and accurately identify excitation surge currents in response to the above technical problems, which will help improve the sensitivity and reliability of the excitation surge current protection device.

[0006] In a first aspect, the present application provides a method for detecting an excitation inrush current, comprising:

[0007] Collect the current signal of the transformer port;

[0008] Determining time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal;

[0009] Determining the time domain characteristic parameters, the frequency domain characteristic parameters, and the importance parameters and correlation parameters of the time-frequency domain characteristic parameters;

[0010] Selecting target feature parameters from the time domain feature parameters, the frequency domain feature parameters, and the time-frequency domain feature parameters according to the importance parameter and the correlation parameter, and constructing a target feature vector of the current signal using the target feature parameters;

[0011] A pre-trained classification model is used to generate a magnetizing inrush current detection result of the transformer port based on a comparison result of the target feature vector and a pre-stored magnetizing inrush current feature.

[0012] In one embodiment, determining the time domain characteristic parameter, the frequency domain characteristic parameter, and the importance parameter and correlation parameter of the time-frequency domain characteristic parameter includes:

[0013] Normalizing the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters to obtain the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters under the same characteristic scale;

[0014] Principal component analysis is performed on the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters under the same characteristic scale to obtain the importance parameter and the correlation parameter.

[0015] In one embodiment, the time domain characteristic parameters include inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters and waveform sharpness characteristic parameters;

[0016] Determining the time domain characteristic parameters of the current signal includes:

[0017] determining an inrush current amplitude characteristic parameter of the current signal according to a target peak value of the current signal;

[0018] determining a current change characteristic parameter of the current signal according to a rising slope of the current signal at an initial stage of a sudden change;

[0019] determining a waveform asymmetry characteristic parameter of the current signal according to the skewness of the current signal;

[0020] A waveform sharpness characteristic parameter of the current signal is determined according to the kurtosis parameter of the current signal.

[0021] In one embodiment, the frequency domain characteristic parameters include fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters and distortion parameters;

[0022] Determining the frequency domain characteristic parameters of the current signal includes:

[0023] Performing Fourier transform processing on the current signal to obtain frequency domain data corresponding to the current signal;

[0024] Extracting fundamental wave amplitude information and each harmonic amplitude information from the frequency domain data, and determining main frequency component parameters in the current signal;

[0025] A distortion parameter corresponding to the current signal is determined according to a total harmonic distortion rate between the fundamental wave amplitude information and the amplitude information of each subharmonic wave.

[0026] In one embodiment, the time-frequency domain characteristic parameters include energy characteristic parameters, energy distribution parameters, main frequency variation parameters and peak variation parameters;

[0027] Determining the time-frequency domain characteristic parameters of the current signal includes:

[0028] performing wavelet decomposition on the current signal to obtain wavelet coefficients at multiple scales;

[0029] Determining energy characteristic parameters of the current signal in frequency bands corresponding to each scale based on energy distribution of wavelet coefficients at each scale;

[0030] determining an energy distribution parameter of the current signal according to entropy values ​​of energy distribution at multiple scales;

[0031] determining, according to the instantaneous frequency trajectory of the current signal, a main frequency variation parameter corresponding to the main frequency variation behavior of the current signal at different time points;

[0032] A peak value variation parameter of the current signal is determined according to a variation of a local energy peak value of the current signal.

[0033] In one embodiment, the method for obtaining the excitation inrush current characteristics includes:

[0034] Collect the original current signal of the transformer port at multiple voltage levels;

[0035] Performing wavelet denoising processing on the original current signal at each voltage level to obtain the denoised original current signal at each voltage level;

[0036] Using a preset bandpass filter to filter the de-noised original current signal to obtain current signal samples at each voltage level;

[0037] The excitation inrush current characteristics at each voltage level are determined according to the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters of the current signal samples at each voltage level.

[0038] In a second aspect, the present application further provides an excitation inrush current detection device, comprising:

[0039] A signal acquisition module is used to collect the current signal of the transformer port;

[0040] A feature extraction module, used to determine the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters of the current signal;

[0041] A feature analysis module, configured to determine the time domain feature parameters, the frequency domain feature parameters, and the importance parameters and correlation parameters of the time-frequency domain feature parameters;

[0042] a vector generation module, configured to select a target characteristic parameter from the time domain characteristic parameter, the frequency domain characteristic parameter, and the time-frequency domain characteristic parameter according to the importance parameter and the correlation parameter, and construct a target characteristic vector of the current signal using the target characteristic parameter;

[0043] The feature comparison module is used to generate an excitation inrush current detection result of the transformer port based on the comparison result of the target feature vector and the pre-stored excitation inrush current feature using a pre-trained classification model.

[0044] 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 the processor implements the method for detecting excitation inrush current according to any one of the embodiments of the first aspect when executing the computer program.

[0045] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for detecting magnetizing inrush current according to any one of the embodiments of the first aspect.

[0046] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the method for detecting magnetizing inrush current according to any one of the embodiments of the first aspect.

[0047] The above-mentioned excitation inrush current detection method, device, computer equipment, computer-readable storage medium and computer program product collect the current signal of the transformer port, extract the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters of the current signal, select target characteristic parameters based on the importance and correlation of the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters to construct a target characteristic vector, and use a classification model based on machine learning to generate the excitation inrush current detection result of the transformer port according to the comparison result of the target characteristic vector and the excitation inrush current characteristic. It can not only make full use of the multi-dimensional characteristics of the current signal in the time domain, frequency domain and time-frequency domain to improve the accuracy of the excitation inrush current detection, but also use the importance and correlation of each characteristic parameter to select the target characteristic parameter for feature dimensionality reduction, which not only retains the original key difference information of the current signal, but also reduces redundant characteristic parameters, thereby helping to improve the computational efficiency and stability of the excitation inrush current detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 A diagram showing an application environment of a method for detecting an excitation inrush current according to an embodiment;

[0050] Figure 2 1 is a flow chart of a method for detecting an excitation inrush current according to an embodiment;

[0051] Figure 3 Schematic diagram of a flow chart of a step of determining time domain characteristic parameters in one embodiment;

[0052] Figure 4 Schematic diagram of a flow chart of a frequency domain characteristic parameter determination step in one embodiment;

[0053] Figure 5 Schematic diagram of a flow chart of a step of determining time-frequency domain characteristic parameters in one embodiment;

[0054] Figure 6 Schematic diagram of a flow chart of a step of determining a magnetizing inrush current characteristic in one embodiment;

[0055] Figure 7 1 is a flow chart of a method for detecting an excitation inrush current according to another embodiment;

[0056] Figure 8 FIG. 8 is a structural block diagram of an excitation inrush current detection device 800 according to an embodiment;

[0057] Figure 9 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0058] 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.

[0059] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0060] The excitation inrush current detection method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown, the computer device 102 can communicate data with the sampling device 104 through a network, a bus or a port. The sampling device 104 can be connected to the transformer 106 through a bus or a port to collect the current signal of the transformer 106 port.

[0061] For example, the computer device 102 can obtain the current signal from the port of the transformer 106 through the sampling device 104. The computer device 102 can use the preset and stored time domain feature extraction logic to perform time domain analysis on the current signal to determine the time domain feature parameters of the current signal. Similarly, the computer device 102 can use the frequency domain feature extraction logic to perform frequency domain analysis on the current signal to obtain the frequency domain feature parameters of the current signal; the computer device 102 can use the time-frequency domain feature extraction logic to perform time-frequency domain analysis on the current signal to obtain the time-frequency domain feature parameters of the current signal.

[0062] The computer device 102 may be deployed with a pre-trained classification model. The classification model may be a machine learning model trained using the magnetizing inrush current characteristics and short-circuit fault current characteristics of a transformer at multiple voltage levels. The computer device 102 may perform an importance assessment on the time-domain characteristic parameters, frequency-domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal, and calculate the correlation between the characteristic parameters to obtain importance parameters and correlation parameters corresponding to the time-domain characteristic parameters, frequency-domain characteristic parameters, and time-frequency domain characteristic parameters. Based on the importance parameters and correlation parameters, target characteristic parameters are extracted from the time-domain characteristic parameters, frequency-domain characteristic parameters, and time-frequency domain characteristic parameters. The target characteristic parameters are used to construct a target feature vector of the current signal. The target feature vector is input into the pre-trained classification model, and the classification model is used to generate a magnetizing inrush current detection result for the transformer port based on a comparison result between the target feature vector and pre-stored magnetizing inrush current characteristics.

[0063] The computer device 102 may include, but is not limited to, a terminal, a server, or a system consisting of a terminal and a server, and may be implemented through interaction between the terminal and the server. Terminals may include, but are not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices may include smart TVs, smart car devices, projectors, etc. Portable wearable devices may include smart watches, smart bracelets, head-mounted devices, etc. Head-mounted devices may include virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server may be a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing cloud computing services. The sampling device 104 may be a high-frequency sampling device with a sampling frequency above 10 kHz to fully preserve the high-frequency dynamic characteristics of the magnetizing inrush current. The transformer 106 may be an electrical energy transmission and conversion device capable of any voltage level.

[0064] In an exemplary embodiment, Figure 2 As shown, a method for detecting excitation inrush current is provided, which is applied to Figure 1 Taking the computer device 102 in the example as an example, the method includes the following steps S202 to S210.

[0065] Step S202: collecting the current signal of the transformer port.

[0066] For example, the computer device may continuously collect current signals from the primary or secondary ports of the transformer using a high-frequency sampling device to obtain the current signals at the transformer ports. Optionally, in some embodiments, the computer device may further perform denoising and filtering on the collected raw signals, and use the processed signals for calculations in subsequent steps S204 to S210.

[0067] Step S204 : determining the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal.

[0068] Among them, the time domain characteristic parameters can be used to reflect the characteristics of the current signal in the time domain, such as amplitude characteristics, change intensity characteristics, waveform sharpness characteristics, etc.

[0069] Frequency domain characteristic parameters can be used to reflect the characteristics of the current signal in the frequency domain, such as the main frequency components and the degree of distortion of the harmonics relative to the fundamental wave.

[0070] The time-frequency domain characteristic parameters can be used to reflect the characteristics of the current signal in the time-frequency domain, such as the capacity characteristics of different frequency bands, the main frequency change behavior at different time points, the complexity and disorder of the capacity distribution, and the change of local peak values.

[0071] For example, the computer device can perform time domain feature analysis on the current signal, extract parameters such as the peak value, slope, and kurtosis of the current signal, and obtain time domain feature parameters of the current signal. The computer device can perform Fourier transform processing on the current signal to obtain frequency domain feature parameters of the current signal. The computer device can perform discrete wavelet transform processing on the current signal to obtain time and frequency domain feature parameters of the current signal.

[0072] Step S206: determining the time domain characteristic parameters, the frequency domain characteristic parameters, and the importance parameters and correlation parameters of the time and frequency domain characteristic parameters.

[0073] The importance parameter can be used to characterize the importance of each feature parameter among all feature parameters.

[0074] The correlation parameter can be used to characterize the correlation between two feature parameters.

[0075] For example, the computer device may read the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal at the transformer port. Based on the importance evaluation logic, the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal are respectively evaluated for importance to obtain importance evaluation parameters of the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters. Based on the correlation evaluation logic, the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters are calculated to obtain correlation parameters of the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters.

[0076] The importance evaluation logic can be implemented based on preset importance weights corresponding to the time domain, frequency domain, or time-frequency domain; or it can be implemented by calculating statistical parameters such as the correlation, chi-square distribution, and mutual information between characteristic parameters and the excitation inrush current characteristics; or it can be implemented based on principal component analysis. The correlation evaluation logic can be implemented by calculating the Pearson correlation coefficient or Spearman correlation coefficient between each characteristic parameter, or it can be implemented based on principal component analysis.

[0077] Step S208 : selecting target feature parameters from the time domain feature parameters, the frequency domain feature parameters, and the time-frequency domain feature parameters according to the importance parameter and the correlation parameter, and constructing a target feature vector of the current signal using the target feature parameters.

[0078] The target feature vector can be used to represent the characteristics of the fused current signal in the time domain, frequency domain and time-frequency domain and is a low-dimensional feature vector after compression.

[0079] Exemplarily, the computer device may use the time domain characteristic parameters, frequency domain characteristic parameters or time-frequency domain characteristic parameters whose importance parameters and correlation parameters meet preset thresholds as target characteristic parameters, and perform feature fusion on the target characteristic parameters to obtain the target characteristic vector of the current signal.

[0080] Step S210 : Using a pre-trained classification model, based on a comparison result between a target feature vector and pre-stored magnetizing inrush current features, a magnetizing inrush current detection result of a transformer port is generated.

[0081] Optionally, in some embodiments, the preset classification model can be obtained by training a machine learning classification model using the magnetizing inrush current characteristics of the transformer at different voltage levels using a computer device.

[0082] For example, the computer device may use a classification model to compare a target feature vector of the current signal with the magnetizing inrush current characteristics learned during the training phase and stored in the model's built-in library. If the comparison result between the target feature vector and the magnetizing inrush current characteristics is consistent (a match is considered consistent if the degree of match / similarity exceeds a preset comparison threshold), a magnetizing inrush current detection result for the transformer port is generated indicating the presence of magnetizing inrush current. Otherwise, a magnetizing inrush current detection result for the transformer port is generated indicating the absence of magnetizing inrush current.

[0083] In the above-mentioned excitation inrush current detection method, the current signal of the transformer port is collected to extract the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters of the current signal. Based on the importance and correlation of the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters, the target characteristic parameters are selected to construct a target characteristic vector. The classification model based on machine learning is used to generate the excitation inrush current detection result of the transformer port according to the comparison result of the target characteristic vector and the excitation inrush current characteristic. This method can not only make full use of the multi-dimensional characteristics of the current signal in the time domain, frequency domain and time-frequency domain to improve the accuracy of the excitation inrush current detection, but also use the importance and correlation of each characteristic parameter to select the target characteristic parameters for feature dimensionality reduction, which not only retains the original key difference information of the current signal, but also reduces redundant characteristic parameters, thereby helping to improve the computational efficiency and stability of the excitation inrush current detection.

[0084] In an exemplary embodiment, step S206 includes normalizing the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters to obtain the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale. Principal component analysis is performed on the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale to obtain importance parameters and correlation parameters.

[0085] Optionally, in some embodiments, the computer device may utilize the maximum and minimum values ​​of the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters to convert the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters into values ​​between [0, 1] using a minimum-maximum normalization method, thereby obtaining the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale. Alternatively, in other embodiments, the computer device may also utilize a standard deviation normalization method (Z-Score Normalization, referred to as Z-Score normalization) to convert the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters into a distribution with a mean of 0 and a standard deviation of 1, thereby obtaining the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale.

[0086] Optionally, in some embodiments, the computer device may calculate the variance of the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale, as well as the sum of the variances of the above characteristic parameters, using a principal component analysis method. The ratio of the variance of each characteristic parameter to the sum of the variances is used as the importance parameter of the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters at the same characteristic scale.

[0087] Optionally, in some embodiments, the computer device can determine the correlation parameters of the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters at the same characteristic scale by using the principal component analysis method and the covariance matrix of the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters at the same characteristic scale.

[0088] Optionally, in some embodiments, the computer device may select target feature parameters from the time-domain feature parameters, frequency-domain feature parameters, and time-frequency-domain feature parameters at the same feature scale based on their importance parameters and correlation parameters. The target feature parameters are then used to construct a target feature vector for the current signal. The target feature vector is then input into a pre-trained classification model to obtain a corresponding inrush current detection result.

[0089] In this embodiment, by normalizing the extracted characteristic parameters in the time domain, frequency domain and time-frequency domain, the characteristic scales of the characteristic parameters can be unified, avoiding the influence of different dimensions of the characteristic parameters on subsequent analysis, and evaluating the importance and correlation of each characteristic parameter by applying the principal component analysis method, so that the principal components of each characteristic parameter can be selected based on the principal component analysis results to form a compressed low-dimensional target feature vector, which can help reduce data redundancy and improve the classification accuracy of the subsequent classification model used for excitation inrush current classification, thereby achieving the effect of improving the computational efficiency and stability of subsequent excitation inrush current detection while retaining the original key difference information of the current signal.

[0090] In an exemplary embodiment, the time domain characteristic parameters include inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters and waveform sharpness characteristic parameters. Figure 3 As shown, step S204 may include steps S302 to S308.

[0091] Step S302: determining the inrush current amplitude characteristic parameter of the current signal according to the target peak value of the current signal.

[0092] The target peak value may be the maximum instantaneous value reached by the signal within a time or space range, and is used to describe the signal strength or amplitude.

[0093] For example, the computer device may extract the maximum peak value from the current signal curve as the target peak value corresponding to the current signal. Alternatively, in some embodiments, the computer device may directly use the target peak value as an inrush current amplitude characteristic parameter reflecting the inrush current amplitude characteristics of the current signal in the time domain. Alternatively, the computer device may utilize the matching relationship between the target peak value and a preset numerical range, and use the amplitude level corresponding to the currently matched preset numerical range as the inrush current amplitude characteristic parameter.

[0094] Step S304 : determining a current change characteristic parameter of the current signal according to the rising slope of the current signal at the initial stage of the sudden change.

[0095] The rise slope (Rise Ramp or Slew Rate) can be a physical quantity that describes the rate of change of a signal when it rises from a low level (such as the minimum value) to a high level (such as the maximum value). It is usually used to analyze the dynamic characteristics of a signal.

[0096] Exemplarily, the computer device can determine the moment when the current signal undergoes a sudden change and the time threshold corresponding to the starting stage to determine the sudden change starting stage of the current signal. Alternatively, the computer device can also use the stage where the rate of change exceeds the preset change threshold as the sudden change starting stage of the current signal based on the rate of change of the current signal. Calculate the ratio between the current signal change amount and the time change amount at the sudden change starting stage, or the derivative of the current signal change amount with respect to the time change amount, as the rising slope of the current signal. Optionally, in some embodiments, the computer device can directly use the rising slope of the current signal at the sudden change starting stage as a current change characteristic parameter to measure the severity of the current change. Alternatively, it is also possible to refer to the above-mentioned method for determining the amplitude level to obtain the degree of change corresponding to the rising slope as the corresponding current change characteristic parameter.

[0097] Step S306 : determining waveform asymmetry characteristic parameters of the current signal according to the slope of the current signal.

[0098] Among them, skewness is a statistic that describes the symmetry of signal distribution and is used to measure the degree to which the signal waveform deviates from the symmetrical distribution.

[0099] For example, the computer device can calculate the skewness of the current signal using the value, mean, and standard deviation of the current signal. The skewness can be directly used as a waveform asymmetry characteristic parameter to measure the asymmetry of the current waveform. Alternatively, the corresponding asymmetry level can be determined based on the numerical range of the skewness as the waveform asymmetry characteristic parameter.

[0100] Step S308 : determining a waveform sharpness characteristic parameter of the current signal according to the kurtosis parameter of the current signal.

[0101] The kurtosis parameter may be a statistic describing the peak degree of the signal distribution shape, and is used to measure the steepness of the signal probability density function compared with the normal distribution.

[0102] For example, the computer device can calculate a kurtosis parameter of the current signal using the value, mean, and standard deviation of the current signal. The kurtosis parameter can be directly used as a waveform sharpness characteristic parameter to measure the sharpness characteristic of the current waveform. Alternatively, a corresponding sharpness level can be determined based on the numerical range of the kurtosis parameter as the waveform sharpness characteristic parameter.

[0103] For example, the computer device may use the inrush current amplitude characteristic parameter, the current change characteristic parameter, the waveform asymmetry characteristic parameter, and the waveform sharpness characteristic parameter determined in steps S302 to S308 as the time domain characteristic parameters of the current signal.

[0104] In this embodiment, by determining the inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters and waveform sharpness characteristic parameters as time domain characteristic parameters, the breadth of the time domain characteristic analysis of the current signal can be improved and the data richness of the time domain characteristic parameters can be improved.

[0105] In an exemplary embodiment, the frequency domain characteristic parameters include fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters and distortion parameters. Figure 4 As shown, step S204 may include steps S402 to S406.

[0106] Step S402 : Perform Fourier transform processing on the current signal to obtain frequency domain data corresponding to the current signal.

[0107] Step S404 : extracting fundamental wave amplitude information and each harmonic amplitude information from the frequency domain data, and determining the main frequency component parameters in the current signal.

[0108] For example, a computer device can perform a Fourier transform on the current signal and use the resulting data as frequency domain data corresponding to the current signal. Based on a preset frequency, fundamental wave amplitude information and each harmonic amplitude information are extracted from the frequency domain data. For example, if the fundamental wave is 50 Hz, the frequency of the first harmonic obtained by Fourier transform may be 100 Hz, the frequency of the second harmonic may be 150 Hz, and the frequency of the third harmonic may be 250 Hz. The main frequency component parameters in the current signal are determined based on the signal frequency corresponding to the peak with the highest amplitude.

[0109] Step S406 : determining a distortion parameter corresponding to the current signal according to the total harmonic distortion ratio between the fundamental wave amplitude information and the amplitude information of each subharmonic wave.

[0110] For example, the computer device can perform a square root operation on the sum of the squares of the harmonic amplitudes and calculate the ratio of the resulting square root to the fundamental amplitude as the total harmonic distortion (THD). This quantitatively describes the overall distortion of the harmonic energy relative to the fundamental energy in the current signal spectrum, serving as an important frequency-domain indicator for distinguishing magnetizing inrush current from short-circuit fault current. Alternatively, the THD can be used directly as a distortion parameter, or a corresponding distortion level can be determined based on the numerical range of the THD as a distortion parameter.

[0111] Optionally, in some implementations, the total harmonic distortion rate may be calculated using the following formula:

[0112] ,

[0113] in, Indicates the total harmonic distortion. i indicates the harmonic order. n indicates the total harmonic order. Indicates the current amplitude of the i-th harmonic. Indicates the fundamental current amplitude.

[0114] For example, the computer device may use the fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters, and distortion parameters determined in steps S402 to S406 as frequency domain characteristic parameters of the current signal.

[0115] In this embodiment, by performing spectrum analysis on the current signal, extracting the fundamental amplitude information, each harmonic amplitude information, main frequency components, and distortion parameters to determine the frequency domain characteristic parameters of the current signal, it can help improve the subsequent classification accuracy of the excitation inrush current and short-circuit fault current.

[0116] In an exemplary embodiment, the time-frequency domain characteristic parameters include energy characteristic parameters, energy distribution parameters, main frequency variation parameters and peak variation parameters. Figure 5 As shown, step S204 may include steps S502 to S510.

[0117] Step S502 : performing wavelet decomposition on the current signal to obtain wavelet coefficients at multiple scales.

[0118] Step S504 : determining energy characteristic parameters of the current signal in frequency bands corresponding to the respective scales according to the energy distribution of the wavelet coefficients at the respective scales.

[0119] Step S506 : determining energy distribution parameters of the current signal according to the entropy values ​​of the energy distribution at multiple scales.

[0120] Among them, each scale corresponds to a frequency band.

[0121] Wavelet coefficients can be used to characterize the numerical results of wavelet transform and to describe the time-frequency characteristics of signals.

[0122] The energy characteristic parameter can be used to reflect the energy characteristics of the current signal in different frequency bands.

[0123] Energy distribution parameters can be used to reflect the complexity and disorder of quantitative energy distribution and assist in identifying the non-stationary characteristics of the excitation inrush current waveform.

[0124] Exemplarily, a computer device can use a discrete wavelet transform method to perform multi-scale wavelet decomposition on the current signal to obtain wavelet coefficients of the current signal at multiple scales. The energy of the wavelet coefficients at each scale is calculated to determine the energy distribution of the wavelet signal at each scale. Based on the energy distribution corresponding to each scale, the degree of energy concentration in the frequency band corresponding to each scale is determined to obtain the energy characteristic parameters of the current signal in the frequency band corresponding to each scale. The entropy value of the energy distribution at multiple scales is obtained by multi-scale wavelet entropy calculation, and the entropy value is directly used as the energy distribution parameter, or the corresponding complex disorder level is determined as the corresponding energy distribution parameter based on the numerical range of the entropy value.

[0125] Optionally, in some embodiments, the computer device may use discrete wavelet transform to decompose the current signal into five scale levels.

[0126] Step S508 : determining the main frequency variation parameters corresponding to the main frequency variation behaviors of the current signal at different time points according to the instantaneous frequency trajectory of the current signal.

[0127] For example, a computer device can use the Hilbert-Huang transform to calculate the instantaneous frequency of the current signal, combining the instantaneous frequencies at different time points to form an instantaneous frequency trajectory of the current signal. Based on the frequency changes in the instantaneous frequency trajectory, the main frequency change behavior of the current signal at different time points is tracked to obtain the main frequency change parameters of the current signal at different time points.

[0128] Step S510 : determining a peak value variation parameter of the current signal according to a variation of a local energy peak value of the current signal.

[0129] Among them, the peak change parameter can be used to provide support for the initial identification of the excitation inrush current.

[0130] For example, the computer device can determine the peak value of the energy distribution at each scale, and determine the local energy peak change corresponding to each scale based on the peak value at each scale, such as the change amount and change amplitude, to obtain the peak change parameter of the current signal.

[0131] For example, the computer device may use the energy characteristic parameters, energy distribution parameters, main frequency variation parameters, and peak value variation parameters determined in steps S504 to S510 as the time-frequency domain characteristic parameters of the current signal.

[0132] In this embodiment, the wavelet coefficients of the current signal at multiple scales are determined by multi-scale wavelet decomposition, the energy distribution, entropy value and energy peak change of the current signal in different frequency bands are calculated, and the instantaneous frequency trajectory of the current signal is used to track the main frequency change behavior of the current signal at different time points to obtain the time-frequency domain characteristic parameters of the current signal, which can capture the dynamic evolution characteristics of the current signal in both time and frequency domains.

[0133] In an exemplary embodiment, Figure 6 As shown, a method for determining the characteristics of an excitation inrush current is also provided, which may include steps S602 to S608.

[0134] Step S602: collecting original current signals of transformer ports at multiple voltage levels.

[0135] Step S604 , performing wavelet denoising processing on the original current signal at each voltage level to obtain the denoised original current signal at each voltage level.

[0136] Step S606 , using a preset bandpass filter to filter the denoised original current signal to obtain current signal samples at various voltage levels.

[0137] Step S608 : determining the magnetizing inrush current characteristics at each voltage level according to the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal samples at each voltage level.

[0138] For example, a computer device can select a high-frequency sampling device with a sampling frequency of 10 kHz or higher to continuously sample the primary or secondary ports of a transformer at multiple voltage levels, obtaining raw current signals at multiple voltage levels. The following operations are performed on the raw current signals at each voltage level: Wavelet denoising is performed on the raw current signals using a preset wavelet basis (e.g., a Daubechies 4th-order wavelet basis) to remove high-frequency interference and power frequency noise from the raw current signals. The denoised raw current signals are filtered using a preset bandpass filter (e.g., a filter with a signal bandwidth limited to 1 Hz to 1000 Hz) to remove low-frequency environmental noise and ultra-high-frequency resonant signals from the transformer system, thereby obtaining the current signal at the current voltage level. Referring to the time domain analysis method, frequency domain analysis method, and time-frequency domain analysis method provided in the above embodiments, the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal samples at the current voltage level are obtained.

[0139] Optionally, in some embodiments, the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters of the current signal samples at each voltage level can be directly used to construct the characteristic vectors of the current signal samples at each voltage level as the excitation inrush current characteristics at each voltage level.

[0140] Alternatively, in other embodiments, the target feature vector construction method described above may be referred to: target feature parameters of the current signal sample are selected based on the time domain feature parameters, frequency domain feature parameters, and importance parameters and correlation parameters of the time-frequency domain feature parameters, and the target feature parameters are used to construct the target feature vector of the current signal sample at the current voltage level. The target feature vectors of the current signal samples at each voltage level are summarized as the magnetizing inrush current feature to construct a feature vector library.

[0141] Optionally, in some implementations, a feature vector library may be used to train a classification model based on machine learning to obtain a pre-trained classification model for generating the excitation inrush current detection results in the above embodiments.

[0142] In this embodiment, wavelet denoising and bandpass filtering are performed on the raw current signal to improve its quality, providing a clean and accurate signal foundation for subsequent feature extraction. Extracting magnetizing inrush current features by acquiring raw current signals at multiple voltage levels enriches the data volume of magnetizing inrush current features, thereby improving the classification accuracy of the subsequently trained classification model.

[0143] In an exemplary embodiment, Figure 7 As shown, a method for detecting an excitation inrush current is also provided, comprising the following steps S702 to S710. In which:

[0144] Step S702 : collecting the real-time current signal of the transformer port, performing wavelet denoising and bandpass filtering on the real-time current signal to obtain the current signal.

[0145] For example, the computer device may use a high-frequency sampling device to obtain a real-time current signal from a transformer port, and perform wavelet denoising and bandpass filtering on the real-time current signal in the manner described in steps S602 to S606 to obtain a current signal.

[0146] Step S704 : determining the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal.

[0147] For example, the computer device may perform time domain feature analysis on the current signal in accordance with the method of steps S302 to S308, and determine the inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters, and waveform sharpness characteristic parameters as time domain feature parameters. A fast Fourier transform may be used to perform spectrum analysis on the current signal in accordance with the method of steps S402 to S406, and determine the fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters, and distortion parameters as frequency domain feature parameters. A wavelet analysis technique may be used to extract time and frequency domain features on the current signal in accordance with the method of steps S502 to S510, and determine the energy characteristic parameters, energy distribution parameters, main frequency change parameters, and peak change parameters as time and frequency domain feature parameters.

[0148] Step S706 , normalizing the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters, and using principal component analysis to determine the importance parameters and correlation parameters of the normalized time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters.

[0149] Exemplarily, the computer device can refer to the normalization method and principal component analysis method provided in the above embodiment to normalize the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters determined in step S704, and use the principal component analysis method to determine the importance parameters and correlation parameters of the normalized time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters.

[0150] Step S708 : selecting target feature parameters from the time domain feature parameters, the frequency domain feature parameters, and the time-frequency domain feature parameters according to the importance parameter and the correlation parameter, and constructing a target feature vector of the current signal using the target feature parameters.

[0151] For example, based on the importance parameter and the correlation parameter determined in step S706, the computer device may select, as the target feature parameter, a time domain feature parameter, a frequency domain feature parameter, or a time-frequency domain feature parameter whose importance parameter is greater than a preset importance threshold and / or whose correlation parameter is greater than a preset correlation threshold. The target feature parameter is then fused to obtain a target feature vector of the current signal.

[0152] Step S710 : Using a pre-trained classification model, a magnetizing inrush current detection result of a transformer port is generated based on a comparison result between a target feature vector and pre-stored magnetizing inrush current features.

[0153] For example, the computer device can use a pre-trained classification model to compare the target feature vector generated in step S708 with pre-stored magnetizing inrush current characteristics to determine whether the current signal meets the magnetizing inrush current characteristics, and generate a magnetizing inrush current detection result for the transformer port. If the magnetizing inrush current detection result indicates the presence of magnetizing inrush current, the computer device can subsequently control the transformer's differential protection to not operate.

[0154] Optionally, in some embodiments, during the classification model training phase, the computer device may perform the following operations for transformers of multiple voltage levels: using the current signal at the transformer port at the current voltage level, performing time-domain, frequency-domain, and time-frequency-domain analysis on the current signal to obtain time-domain characteristic parameters, frequency-domain characteristic parameters, and time-frequency-domain characteristic parameters of the current signal; performing feature fusion and dimensionality reduction on the time-domain characteristic parameters, frequency-domain characteristic parameters, and time-frequency-domain characteristic parameters based on importance and relevance to obtain target characteristic parameters; and using the target characteristic parameters to construct a magnetizing inrush current feature vector for the current voltage level. A feature vector library is constructed using the magnetizing inrush current feature vectors for multiple voltage levels. Data from the feature vector library is input into an initial machine learning classification model for training and testing, resulting in a classification model for use in the application phase. The classification model is then connected to the power grid system, enabling subsequent classification models to generate magnetizing inrush current detection results for the transformer port based on the target characteristic vectors of the current signal read in real time from the transformer port. Because the machine learning classification model has good generalization capabilities and high classification accuracy, it can achieve real-time differentiation between magnetizing inrush current and short-circuit fault current.

[0155] In this embodiment, by performing wavelet denoising and bandpass filtering on the original current signal, high-frequency interference and power frequency noise can be removed, environmental low-frequency noise and system ultra-high frequency resonance signals can be filtered out, and the quality of the current signal can be improved. By performing feature analysis and extraction on the current signal in the time domain, frequency domain, and time-frequency domain, typical characteristics of the current signal in the time domain, frequency domain, and time-frequency domain are obtained, and the importance and correlation of each feature are evaluated based on the principal component analysis method to achieve effective feature fusion and dimensionality reduction processing. While retaining the key difference information of the original signal, the computational efficiency and stability of the subsequent recognition algorithm are improved. By adopting a machine learning-based classification model to perform excitation inrush current detection according to the target feature vector, the multi-dimensional characteristics of the excitation inrush current in the time domain, frequency domain, and time-frequency domain can be fully utilized to improve the recognition accuracy and speed of the excitation inrush current, and achieve high-precision, high-robustness, and high-real-time excitation inrush current recognition to meet the future smart grid's demand for intelligent and sensitive development of transformer protection equipment.

[0156] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence as indicated by the arrows, these steps are not necessarily executed 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 may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times. 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 the steps or stages in other steps.

[0157] Based on the same inventive concept, embodiments of the present application further provide an excitation inrush current detection device for implementing the aforementioned excitation inrush current detection method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more of the following embodiments of the excitation inrush current detection device can be found in the aforementioned limitations of the excitation inrush current detection method and are not further elaborated here.

[0158] In an exemplary embodiment, Figure 8 As shown, an excitation inrush current detection device 800 is provided, comprising: a signal acquisition module 802, a feature extraction module 804, a feature analysis module 806, a vector generation module 808 and a feature comparison module 810, wherein:

[0159] The signal acquisition module 802 is used to acquire the current signal of the transformer port.

[0160] The feature extraction module 804 is used to determine the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters of the current signal.

[0161] The feature analysis module 806 is used to determine the time domain feature parameters, the frequency domain feature parameters, and the importance parameters and correlation parameters of the time and frequency domain feature parameters.

[0162] The vector generation module 808 is used to select target feature parameters from the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters according to the importance parameter and the correlation parameter, and construct a target feature vector of the current signal using the target feature parameters.

[0163] The feature comparison module 810 is configured to generate a magnetizing inrush current detection result of the transformer port based on a comparison result of a target feature vector and pre-stored magnetizing inrush current features using a pre-trained classification model.

[0164] In an exemplary embodiment, the feature analysis module 806 is also used to normalize the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters to obtain the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters at the same feature scale; and perform principal component analysis on the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters at the same feature scale to obtain importance parameters and correlation parameters.

[0165] In an exemplary embodiment, the feature extraction module 804 is also used to determine the inrush current amplitude characteristic parameters of the current signal based on the target peak value of the current signal; determine the current change characteristic parameters of the current signal based on the rising slope of the current signal in the initial stage of the mutation; determine the waveform asymmetry characteristic parameters of the current signal based on the skewness of the current signal; determine the waveform sharpness characteristic parameters of the current signal based on the kurtosis parameter of the current signal; and use the inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters and waveform sharpness characteristic parameters as time domain characteristic parameters.

[0166] In an exemplary embodiment, the feature extraction module 804 is also used to perform Fourier transform processing on the current signal to obtain frequency domain data corresponding to the current signal; extract fundamental wave amplitude information and each harmonic amplitude information from the frequency domain data, and determine the main frequency component parameters in the current signal; determine the distortion parameters corresponding to the current signal based on the total harmonic distortion rate between the fundamental wave amplitude information and each harmonic amplitude information; and use the fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters and distortion parameters as frequency domain feature parameters.

[0167] In an exemplary embodiment, the feature extraction module 804 is also used to perform wavelet decomposition on the current signal to obtain wavelet coefficients at multiple scales; determine the energy characteristic parameters of the current signal in the frequency band corresponding to each scale based on the energy distribution of the wavelet coefficients at each scale; determine the energy distribution parameters of the current signal based on the entropy value of the energy distribution at multiple scales; determine the main frequency change parameters corresponding to the main frequency change behavior of the current signal at different time points based on the instantaneous frequency trajectory of the current signal; determine the peak change parameters of the current signal based on the change of the local energy peak of the current signal; and use the energy characteristic parameters, energy distribution parameters, main frequency change parameters and peak change parameters as time-frequency domain feature parameters.

[0168] In an exemplary embodiment, the signal acquisition module 802 is further configured to acquire raw current signals from transformer ports at multiple voltage levels. The excitation inrush current detection device 800 further includes a model training module configured to perform wavelet denoising on the raw current signals at each voltage level to obtain denoised raw current signals at each voltage level; filter the denoised raw current signals using a preset bandpass filter to obtain current signal samples at each voltage level; and determine the excitation inrush current characteristics at each voltage level based on the time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal samples at each voltage level.

[0169] Each module in the excitation inrush current detection device 800 may be implemented in whole or in part by software, hardware, or a combination thereof. Each module may be embedded in or independent of a processor in a computer device in the form of hardware, or may 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.

[0170] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 9 As shown. The computer device includes a processor, a memory, an input / output (I / O) interface, and a communication interface. The processor, memory, and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. 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, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data such as current signals, time-domain characteristic parameters, frequency-domain characteristic parameters, time-frequency domain characteristic parameters, importance parameters, correlation parameters, target characteristic parameters, target characteristic vectors, excitation inrush current characteristics, comparison results, and excitation inrush current detection results. The I / O interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor, the computer program implements an excitation inrush current detection method.

[0171] Those skilled in the art will understand that Figure 9 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.

[0172] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0173] In an exemplary 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 steps in the above-mentioned method embodiments are implemented.

[0174] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0175] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. 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 above-mentioned embodiments. In particular, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of a non-volatile memory and a volatile memory. Non-volatile memory can 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 can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases 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, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0176] The technical features of the above embodiments can be combined arbitrarily. In order 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 application.

[0177] 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 detecting an excitation inrush current, characterized in that: The method comprises: Collect the current signal of the transformer port; Determining time domain characteristic parameters, frequency domain characteristic parameters, and time-frequency domain characteristic parameters of the current signal; Determining the time domain characteristic parameters, the frequency domain characteristic parameters, and the importance parameters and correlation parameters of the time-frequency domain characteristic parameters; Selecting target feature parameters from the time domain feature parameters, the frequency domain feature parameters, and the time-frequency domain feature parameters according to the importance parameter and the correlation parameter, and constructing a target feature vector of the current signal using the target feature parameters; A pre-trained classification model is used to generate a magnetizing inrush current detection result of the transformer port based on a comparison result of the target feature vector and a pre-stored magnetizing inrush current feature.

2. The method according to claim 1, characterized in that The determining of the time domain characteristic parameter, the frequency domain characteristic parameter, and the importance parameter and the correlation parameter of the time-frequency domain characteristic parameter includes: Normalizing the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters to obtain the time domain characteristic parameters, the frequency domain characteristic parameters, and the time-frequency domain characteristic parameters under the same characteristic scale; Principal component analysis is performed on the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters under the same characteristic scale to obtain the importance parameter and the correlation parameter.

3. The method according to claim 1, characterized in that The time domain characteristic parameters include inrush current amplitude characteristic parameters, current change characteristic parameters, waveform asymmetry characteristic parameters and waveform sharpness characteristic parameters; Determining the time domain characteristic parameters of the current signal includes: determining an inrush current amplitude characteristic parameter of the current signal according to a target peak value of the current signal; determining a current change characteristic parameter of the current signal according to a rising slope of the current signal at an initial stage of a sudden change; determining a waveform asymmetry characteristic parameter of the current signal according to the skewness of the current signal; A waveform sharpness characteristic parameter of the current signal is determined according to the kurtosis parameter of the current signal.

4. The method according to claim 1, wherein The frequency domain characteristic parameters include fundamental wave amplitude information, each harmonic amplitude information, main frequency component parameters and distortion parameters; Determining the frequency domain characteristic parameters of the current signal includes: Performing Fourier transform processing on the current signal to obtain frequency domain data corresponding to the current signal; Extracting fundamental wave amplitude information and each harmonic amplitude information from the frequency domain data, and determining main frequency component parameters in the current signal; A distortion parameter corresponding to the current signal is determined according to a total harmonic distortion rate between the fundamental wave amplitude information and the amplitude information of each subharmonic wave.

5. The method according to claim 1, wherein The time-frequency domain characteristic parameters include energy characteristic parameters, energy distribution parameters, main frequency variation parameters and peak value variation parameters; Determining the time-frequency domain characteristic parameters of the current signal includes: performing wavelet decomposition on the current signal to obtain wavelet coefficients at multiple scales; Determining energy characteristic parameters of the current signal in frequency bands corresponding to each scale based on energy distribution of wavelet coefficients at each scale; determining an energy distribution parameter of the current signal according to entropy values ​​of energy distribution at multiple scales; determining, according to the instantaneous frequency trajectory of the current signal, a main frequency variation parameter corresponding to the main frequency variation behavior of the current signal at different time points; A peak value variation parameter of the current signal is determined according to a variation of a local energy peak value of the current signal.

6. The method according to any one of claims 1 to 5, characterized in that The method for obtaining the excitation inrush current characteristics includes: Collect the original current signal of the transformer port at multiple voltage levels; Performing wavelet denoising processing on the original current signal at each voltage level to obtain the denoised original current signal at each voltage level; Using a preset bandpass filter to filter the de-noised original current signal to obtain current signal samples at each voltage level; The excitation inrush current characteristics at each voltage level are determined according to the time domain characteristic parameters, frequency domain characteristic parameters and time-frequency domain characteristic parameters of the current signal samples at each voltage level.

7. An excitation inrush current detection device, characterized in that: The device comprises: A signal acquisition module is used to collect the current signal of the transformer port; A feature extraction module, used to determine the time domain feature parameters, frequency domain feature parameters and time-frequency domain feature parameters of the current signal; A feature analysis module, configured to determine the time domain feature parameters, the frequency domain feature parameters, and the importance parameters and correlation parameters of the time-frequency domain feature parameters; a vector generation module, configured to select a target characteristic parameter from the time domain characteristic parameter, the frequency domain characteristic parameter, and the time-frequency domain characteristic parameter according to the importance parameter and the correlation parameter, and construct a target characteristic vector of the current signal using the target characteristic parameter; The feature comparison module is used to generate an excitation inrush current detection result of the transformer port based on the comparison result of the target feature vector and the pre-stored excitation inrush current feature using a pre-trained classification model.

8. 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 6 are implemented.

9. 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 6 are implemented.

10. 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 6 are implemented.

Images