Power distribution equipment self-checking method and system based on intelligent circuit breaker

Abstract

The invention provides a power distribution equipment self-checking method and system based on an intelligent circuit breaker, and the method comprises the steps: calculating all-phase frequency spectrums of current data of two continuous full-period alternating currents, and obtaining a differential frequency spectrum matrix through calculation based on the two all-phase frequency spectrums; current data of the full-period alternating current corresponding to the all-phase frequency spectrum with the maximum frequency spectrum energy serves as target current data, a time-frequency two-dimensional energy distribution diagram of the target current data is generated, and a time-frequency information entropy value is obtained through calculation based on the time-frequency two-dimensional energy distribution diagram; obtaining a differential frequency spectrum matrix sequence and a time-frequency two-dimensional energy distribution diagram sequence according to a preset sliding window length and step length, and extracting features of the differential frequency spectrum matrix sequence and the time-frequency two-dimensional energy distribution diagram sequence; and determining a fusion weight based on the fundamental wave amplitude of the current data in the sliding window and the total harmonic distortion degree, carrying out weighted fusion on the features of the differential frequency spectrum matrix sequence and the features of the time-frequency two-dimensional energy distribution diagram sequence by using the fusion weight, and obtaining a self-inspection result based on a weighted fusion result.

Description

Technical Field

[0001] This application belongs to the field of equipment self - inspection, and particularly relates to a self - inspection method and system for distribution equipment based on an intelligent circuit breaker. Background Technique

[0002] Traditional distribution equipment protection devices such as fuses and thermal - magnetic circuit breakers mainly provide passive protection against strong faults such as overload and short - circuit. With the development of smart grid technology, self - inspection technology can upgrade equipment protection from passive response to active warning. Through built - in sensors and microprocessors, an intelligent circuit breaker can continuously monitor key parameters such as current, voltage, power, and temperature of the line in real - time for 7×24 hours. It can not only quickly cut off the circuit when a fault occurs, but also identify potential safety hazards such as loose connections, aging of the insulation layer, and three - phase imbalance in advance by analyzing subtle changes in electrical data, thus avoiding serious accidents such as electrical fires and equipment damage.

[0003] Compared with overload, etc., accurate detection of series arc faults is a difficult point in the self - inspection of distribution equipment by intelligent circuit breakers. Series arcs are caused by reasons such as poor wire contact, loose connection, or insulation damage. Their fault currents are often not sufficient to trigger traditional over - current protection, but the local high temperature generated is sufficient to ignite surrounding combustibles in a short time, which is the primary culprit leading to electrical fires. Arc signals are manifested as wide - spectrum, high - frequency, non - stationary random noise superimposed on the power - frequency current. However, many daily electrical equipment also generates electrical noise highly similar to it during normal operation. For example, commutation sparks generated by series - wound motors such as electric drills and vacuum cleaners, high - frequency switching noise generated by switching power supplies such as calculators and energy - saving lamps, and waveform distortion generated by phase - controlled loads such as dimming lamps. This high similarity between fault characteristics and normal interference leads to the common problems of high false - alarm rate and high miss - alarm rate in existing detection methods. Time - domain analysis methods and frequency - domain analysis methods are difficult to distinguish the random wide - spectrum of arcs from the normal operating spectra of some electrical appliances under complex interference backgrounds and are difficult to cope with complex and changeable actual working conditions. Summary of the Invention

[0004] Aiming at the problem that the arc - fault detection method in the self - inspection of distribution equipment by current intelligent circuit breakers has insufficient ability to identify complex interference sources and is prone to false alarms and miss alarms, this application proposes a self - inspection method for distribution equipment based on an intelligent circuit breaker, including: Obtain the current data of two consecutive full-cycle alternating currents monitored by the intelligent circuit breaker, perform full-phase Fourier transform on the current data of the two consecutive full-cycle alternating currents respectively to obtain two full-phase frequency spectra, and calculate a differential frequency spectrum matrix based on the two full-phase frequency spectra; calculate the spectral energy of the two full-phase frequency spectra within a preset high-frequency band respectively, take the current data of the full-cycle alternating current corresponding to the full-phase frequency spectrum with the largest spectral energy as the target current data, perform continuous wavelet transform on the target current data to generate a time-frequency two-dimensional energy distribution map, and calculate a time-frequency information entropy value based on the time-frequency two-dimensional energy distribution map; Obtain a differential frequency spectrum matrix sequence and a time-frequency two-dimensional energy distribution map sequence according to a preset sliding window length and step size, obtain a time-frequency information entropy value sequence corresponding to the time-frequency two-dimensional energy distribution map sequence, obtain the channel attention of the differential frequency spectrum matrix sequence according to the time-frequency information entropy value sequence, and extract the features of the differential frequency spectrum matrix sequence by using the channel attention; extract the features of the time-frequency two-dimensional energy distribution map sequence based on the stability of the frequency components in the differential frequency spectrum matrix; Determine a fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window, use the fusion weight to perform weighted fusion on the features of the differential frequency spectrum matrix sequence and the features of the time-frequency two-dimensional energy distribution map sequence, and obtain a self-check result based on the result of the weighted fusion.

[0005] Optionally, the obtaining the channel attention of the differential frequency spectrum matrix sequence according to the time-frequency information entropy value sequence includes: Input the time-frequency information entropy value sequence into a multi-layer perceptron, and the multi-layer perceptron includes at least one fully connected layer and one activation layer; Use the multi-layer perceptron to perform a non-linear transformation on the time-frequency information entropy value sequence to obtain an attention weight vector with the same number as the differential frequency spectrum matrix sequence, and each value in the attention weight vector corresponds to a differential frequency spectrum matrix in the differential frequency spectrum matrix sequence.

[0006] Optionally, the obtaining the channel attention of the differential frequency spectrum matrix sequence according to the time-frequency information entropy value sequence includes: Perform normalization processing on each entropy value in the time-frequency information entropy value sequence to obtain a normalized entropy value sequence; Multiply each matrix in the normalized entropy value sequence and the differential frequency spectrum matrix sequence element by element to obtain a weighted differential frequency spectrum matrix sequence; Sum the rows and columns of each matrix in the weighted differential frequency spectrum matrix sequence to obtain a channel attention vector; Normalize the channel attention vector to obtain the channel attention.

[0007] Optionally, extracting the features of the time-frequency two-dimensional energy distribution map sequence based on the stability of the frequency components in the differential frequency spectrum matrix includes: For each time-frequency two-dimensional energy distribution map in the time-frequency two-dimensional energy distribution map sequence, obtain the differential frequency spectrum matrix at the same position in the sliding window; Calculate the absolute value of each frequency component in the differential frequency spectrum matrix to obtain a frequency instability vector; Normalize the frequency instability vector to obtain a frequency attention weight vector; Multiply each row of the time-frequency two-dimensional energy distribution map by the corresponding weight value in the frequency attention weight vector to obtain an enhanced time-frequency two-dimensional energy distribution map; Input the enhanced time-frequency two-dimensional energy distribution map sequence into a two-dimensional convolutional neural network to obtain the features of the time-frequency two-dimensional energy distribution map sequence.

[0008] Optionally, determining the fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window includes: Calculate the fundamental wave amplitude and total harmonic distortion of each current data within the sliding window; perform normalization processing on the fundamental wave amplitude and total harmonic distortion to obtain the normalized fundamental wave amplitude and normalized total harmonic distortion; Input the normalized fundamental wave amplitude and normalized total harmonic distortion into a weighting function to obtain the fusion weight.

[0009] This application also proposes a self-checking system for distribution equipment based on an intelligent circuit breaker, including: A preprocessing unit, configured to obtain the current data of two consecutive full-cycle alternating currents monitored by the intelligent circuit breaker, perform full-phase Fourier transform on the current data of the two consecutive full-cycle alternating currents respectively to obtain two full-phase frequency spectra, calculate a differential frequency spectrum matrix based on the two full-phase frequency spectra; calculate the spectral energy of the two full-phase frequency spectra within a preset high-frequency band respectively, use the current data of the full-cycle alternating current corresponding to the full-phase frequency spectrum with the maximum spectral energy as the target current data, perform continuous wavelet transform on the target current data to generate a time-frequency two-dimensional energy distribution map, and calculate the time-frequency information entropy value based on the time-frequency two-dimensional energy distribution map; A feature extraction unit, configured to obtain a differential frequency spectrum matrix sequence and a time-frequency two-dimensional energy distribution map sequence according to a preset sliding window length and step size, obtain a time-frequency information entropy value sequence corresponding to the time-frequency two-dimensional energy distribution map sequence, obtain the channel attention of the differential frequency spectrum matrix sequence according to the time-frequency information entropy value sequence, and extract the features of the differential frequency spectrum matrix sequence by using the channel attention; extract the features of the time-frequency two-dimensional energy distribution map sequence based on the stability of the frequency components in the differential frequency spectrum matrix; A self-test unit is used to determine a fusion weight based on the fundamental amplitude and total harmonic distortion of the current data in the sliding window, use the fusion weight to weightedly fuse the characteristics of the differential spectrum matrix sequence and the characteristics of the time-frequency two-dimensional energy distribution diagram sequence, and obtain a self-test result based on the result of the weighted fusion.

[0010] Optionally, obtaining the channel attention of the differential spectrum matrix sequence according to the time-frequency information entropy value sequence includes: Inputting the time-frequency information entropy value sequence into a multi-layer perceptron, wherein the multi-layer perceptron includes at least one fully connected layer and one activation layer; The multilayer perceptron is used to perform a nonlinear transformation on the time-frequency information entropy value sequence to obtain attention weight vectors with the same number as the differential spectrum matrix sequence, and each value in the attention weight vector corresponds to a differential spectrum matrix in the differential spectrum matrix sequence.

[0011] Optionally, obtaining the channel attention of the differential spectrum matrix sequence according to the time-frequency information entropy value sequence includes: Normalizing each entropy value in the time-frequency information entropy value sequence to obtain a normalized entropy value sequence; Multiplying the normalized entropy value sequence by each matrix in the differential spectrum matrix sequence element by element to obtain a weighted differential spectrum matrix sequence; Sum the rows and columns of each matrix in the weighted difference spectrum matrix sequence to obtain the channel attention vector; Channel attention is obtained by normalizing the channel attention vector.

[0012] Optionally, the extracting features of the time-frequency two-dimensional energy distribution diagram sequence based on the stability of frequency components in the differential spectrum matrix includes: For each time-frequency two-dimensional energy distribution graph in the time-frequency two-dimensional energy distribution graph sequence, obtaining the differential spectrum matrix at the same position of the sliding window; Calculating the absolute value of each frequency component in the differential spectrum matrix to obtain a frequency instability vector; Normalizing the frequency instability vector to obtain a frequency attention weight vector; Multiplying each row of the time-frequency two-dimensional energy distribution map by the corresponding weight value in the frequency attention weight vector to obtain an enhanced time-frequency two-dimensional energy distribution map; The enhanced time-frequency two-dimensional energy distribution map sequence is input into a two-dimensional convolutional neural network to obtain the characteristics of the time-frequency two-dimensional energy distribution map sequence.

[0013] Optionally, determining the fusion weight based on the fundamental amplitude and total harmonic distortion of the current data in the sliding window includes: Calculate the fundamental wave amplitude and total harmonic distortion of each current data within the sliding window; perform normalization processing on the fundamental wave amplitude and total harmonic distortion to obtain the normalized fundamental wave amplitude and normalized total harmonic distortion; Input the normalized fundamental wave amplitude and normalized total harmonic distortion into the weighting function to obtain the fusion weight.

[0014] The present invention captures the instability of the arc by calculating the differential spectral matrix of consecutive periods, and determines the degree of chaos in combination with time-frequency information entropy, greatly enhancing the recognition ability of fault features; moreover, the present invention adjusts the channel attention using the time-frequency information entropy value to make the extraction of features of the differential spectral matrix sequence more accurate, and uses the stability of the frequency components in the differential spectral matrix to extract the features of the time-frequency two-dimensional energy distribution map sequence to enhance the key frequencies. Further, the real-time working condition of the line, that is, the load level characterized by the fundamental wave amplitude and the power quality characterized by the total harmonic distortion, is used to adjust the fusion weight of the multi-dimensional features, enhancing the accuracy and reliability under different load types and electromagnetic environments, thereby improving the accuracy of arc fault detection and reducing missed reports and false alarms. Brief Description of the Drawings

[0015] Figure 1 is the flowchart of the first embodiment; Figure 2 is the schematic diagram of the current data; Figure 3 is the schematic diagram of the amplitude spectrum and phase spectrum of the all-phase spectrum; Figure 4 is the schematic diagram of the time-frequency two-dimensional energy distribution map; Figure 5 is the schematic diagram of the differential spectral matrix sequence and the time-frequency two-dimensional energy distribution map sequence. Detailed Embodiment

[0016] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0017] Specific embodiments are as follows Figure 1 As shown, the present application proposes a self-checking method for distribution equipment based on an intelligent circuit breaker, including: S1. Obtain the current data of two consecutive full - cycle alternating currents monitored by the intelligent circuit breaker. Perform full - phase Fourier transform on the current data of the two consecutive full - cycle alternating currents respectively to obtain two full - phase frequency spectra. Calculate the differential frequency - spectrum matrix based on the two full - phase frequency spectra. Calculate the spectral energy of the two full - phase frequency spectra within a preset high - frequency band respectively. Take the current data of the full - cycle alternating current corresponding to the full - phase frequency spectrum with the maximum spectral energy as the target current data. Perform continuous wavelet transform on the target current data to generate a time - frequency two - dimensional energy distribution map. Calculate the time - frequency information entropy value based on the time - frequency two - dimensional energy distribution map. The intelligent circuit breaker continuously monitors the current flowing through the line, as Figure 2 shown, and obtains the current data of two closely - connected complete cycles from it. Taking 50Hz alternating current as an example, the current waveform data with a total duration of 40 milliseconds is obtained. Perform full - phase Fourier transform on these two 20 - millisecond current waveforms respectively to obtain two frequency spectra, such as spectrum A and spectrum B. Subtract the value of each frequency point in spectrum B from the corresponding frequency point in spectrum A to obtain the differential frequency spectrum. The differential frequency spectrum highlights the frequency components that change between the two cycles, while the stable background harmonics, etc. will be cancelled out. Spectrum A and spectrum B are both complex - number sequences, assumed to be and respectively. After subtraction, it is . Separate the real part and the imaginary part of the one - dimensional complex differential frequency spectrum, as Figure 3 shown, and arrange them into a two - dimensional matrix. The obtained N×2 two - dimensional matrix is the differential frequency - spectrum matrix. The first column of the differential frequency - spectrum matrix is the real part of the differential complex number at all frequency points, and the second column is the imaginary part of the differential complex number at all frequency points. The differential frequency - spectrum matrix contains the real - part difference representing the change of the waveform in the cosine component and the imaginary - part difference representing the change of the waveform in the sine component.

[0018] Set a high - frequency band, such as 2kHz to 100kHz. Calculate the total energy of spectrum A and spectrum B within this high - frequency band respectively. If the high - frequency energy of spectrum B is much greater than that of spectrum A, since the continuous existence of a dangerous arc will definitely lead to an abnormal increase in the high - frequency - band spectral energy of the line, and without the increase in high - frequency energy, there cannot be a real arc, then the second cycle is more likely to be the cycle when the arc occurs. Select the current data of the second cycle as the target current data. Perform continuous wavelet transform on the target current data to obtain a time - frequency two - dimensional energy distribution map. The horizontal axis of the time - frequency two - dimensional energy distribution map is time, and the vertical axis is frequency. The brightness of each point in the figure represents the energy intensity at that moment and that frequency. Further calculate the overall information entropy of the time - frequency two - dimensional energy distribution map to obtain the time - frequency information entropy value. The time - frequency information entropy value represents the degree of chaos of the energy distribution on the map. A highly chaotic and disordered distribution will obtain a very high entropy value.

[0019] S2. Obtain a sequence of differential frequency spectrum matrices and a sequence of time-frequency two-dimensional energy distribution diagrams according to a preset sliding window length and step size, obtain a sequence of time-frequency information entropy values corresponding to the sequence of time-frequency two-dimensional energy distribution diagrams, obtain the channel attention of the sequence of differential frequency spectrum matrices according to the sequence of time-frequency information entropy values, and extract the features of the sequence of differential frequency spectrum matrices by using the channel attention; extract the features of the sequence of time-frequency two-dimensional energy distribution diagrams based on the stability of the frequency components in the differential frequency spectrum matrix; Obtain a preset sliding window, assume the length is 100 ms and the step size is 20 ms. Each window position contains a sequence composed of 5 groups of initial features, that is, a sequence of 5 differential frequency spectrum matrices, a sequence of 5 time-frequency two-dimensional energy distribution diagrams, and a corresponding sequence of 5 time-frequency information entropy values, as Figure 5 shown.

[0020] For the feature extraction of the sequence of differential frequency spectrum matrices, obtain the channel attention by using the sequence of time-frequency information entropy values. For example, map 5 time-frequency information entropy values to 5 weight values ranging from 0 to 1 through an activation function. Then, extract the respective preliminary features from 5 differential frequency spectrum matrices, and perform weighted summation on these 5 preliminary features according to the weight values to obtain a feature vector as the feature of the sequence of differential frequency spectrum matrices.

[0021] For the feature extraction of the sequence of time-frequency two-dimensional energy distribution diagrams, use the information of the sequence of differential frequency spectrum matrices to enhance the extraction process. For example, when analyzing the 3rd time-frequency two-dimensional energy distribution diagram, obtain the 3rd differential frequency spectrum matrix, and it is calculated that the values of two frequency points, 15 kHz and 40 kHz, in the 3rd differential frequency spectrum matrix are particularly high, indicating that these two frequencies are the most unstable. Then, from the 3rd time-frequency two-dimensional energy distribution diagram, obtain two horizontal row data representing 15 kHz and 40 kHz to get a compressed time-frequency diagram, and input the compressed time-frequency diagram into, for example, a recurrent neural network to obtain the feature of the 3rd time-frequency two-dimensional energy distribution diagram. Perform the above operations on all 5 time-frequency diagrams in the sequence to obtain the features of the sequence of time-frequency two-dimensional energy distribution diagrams.

[0022] S3. Determine the fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window, use the fusion weight to perform weighted fusion on the features of the sequence of differential frequency spectrum matrices and the features of the sequence of time-frequency two-dimensional energy distribution diagrams, and obtain the self-check result based on the result of the weighted fusion.

[0023] Calculate two operating condition indicators, namely the fundamental wave amplitude and the total harmonic distortion, for the current - time - window current data. Assume that these two operating condition indicators indicate that the current line is in a state of high load and serious harmonic pollution. Input the fundamental wave amplitude and the total harmonic distortion, these two operating condition indicators, into a pre - trained decision tree model. The decision tree model will output a set of optimal fusion weights according to the input operating conditions. For example, for the operating condition of high load and high harmonic pollution, the decision tree outputs 0.3 and 0.7. Multiply the differential spectrum matrix sequence features obtained in the second step by 0.3, multiply the time - frequency two - dimensional energy distribution map sequence features by 0.7, and then add the two to obtain a fused feature vector. Input the fused feature vector into a logistic regression classifier. If the confidence level output by the classifier is higher than a preset fault threshold, such as 0.95, for a continuous period of time, it is determined that there is an arc fault in the current distribution equipment, and an alarm signal is generated.

[0024] In an alternative embodiment, obtaining the channel attention of the differential spectrum matrix sequence according to the time - frequency information entropy value sequence includes: Input the time - frequency information entropy value sequence into a multi - layer perceptron, where the multi - layer perceptron includes at least one fully - connected layer and one activation layer; Use the multi - layer perceptron to perform a non - linear transformation on the time - frequency information entropy value sequence to obtain an attention weight vector with the same number of elements as the differential spectrum matrix sequence. Each value in the attention weight vector corresponds to a differential spectrum matrix in the differential spectrum matrix sequence.

[0025] Specifically, assume that within a sliding window, the sequence composed of five time - frequency information entropy values is [1.2, 1.5, 3.8, 1.4, 1.1]. Input the sequence composed of five numbers into a multi - layer perceptron. The fully - connected layer of the multi - layer perceptron performs operations on these five input values, and then the subsequent activation layer performs non - linear processing. After the non - linear transformation of the multi - layer perceptron, a new five - dimensional vector is output, that is, the attention weight vector, such as [0.05, 0.10, 0.70, 0.10, 0.05]. The weight corresponding to the information of the fifth differential spectrum matrix is 0.05, while the weight corresponding to the information of the third matrix is 0.70. Since the time - frequency information entropy value measured at the third time point is the highest, representing the most chaotic signal, the main attention is concentrated on the third differential spectrum matrix, thereby reducing the attention to the matrix information at other more stable time points.

[0026] In an alternative embodiment, obtaining the channel attention of the differential spectrum matrix sequence according to the time - frequency information entropy value sequence includes: Perform normalization processing on each entropy value in the time - frequency information entropy value sequence to obtain a normalized entropy value sequence; Multiplying the normalized entropy value sequence by each matrix in the differential spectrum matrix sequence element by element to obtain a weighted differential spectrum matrix sequence; Sum the rows and columns of each matrix in the weighted difference spectrum matrix sequence to obtain the channel attention vector; Channel attention is obtained by normalizing the channel attention vector.

[0027] Specifically, assume that within a time window, the sequence of five time-frequency information entropy values is [1.2, 1.5, 3.8, 1.4, 1.1]. Normalization converts these values to the range [0, 1], resulting in a sequence of normalized entropy values. The first normalized entropy value is multiplied by each element in the first matrix, the second normalized entropy value is multiplied by each element in the second matrix, and so on. Since the third entropy value, 3.8, has the largest value, its corresponding normalized value is also the largest. Therefore, all values in the third difference spectrum matrix are proportionally amplified, while the values in the other matrices are amplified less. This results in a sequence of weighted difference spectrum matrices, where the overall value of the third matrix is enhanced. For each matrix in the weighted difference spectrum matrix sequence, the values of all its elements are summed, compressing the entire matrix into a single value. The third matrix is amplified the most and has the largest value, resulting in a new vector consisting of five values, which is the channel attention vector. The channel attention vector is normalized again so that the sum of all elements is equal to 1 to obtain the channel attention, which is exemplarily [0.1, 0.15, 0.5, 0.15, 0.1].

[0028] In an optional embodiment, the extracting features of the time-frequency two-dimensional energy distribution diagram sequence based on the stability of the frequency components in the differential spectrum matrix includes: For each time-frequency two-dimensional energy distribution graph in the time-frequency two-dimensional energy distribution graph sequence, obtaining the differential spectrum matrix at the same position of the sliding window; Calculating the absolute value of each frequency component in the differential spectrum matrix to obtain a frequency instability vector; Normalizing the frequency instability vector to obtain a frequency attention weight vector; Multiplying each row of the time-frequency two-dimensional energy distribution map by the corresponding weight value in the frequency attention weight vector to obtain an enhanced time-frequency two-dimensional energy distribution map; The enhanced time-frequency two-dimensional energy distribution map sequence is input into a two-dimensional convolutional neural network to obtain the characteristics of the time-frequency two-dimensional energy distribution map sequence.

[0029] Specifically, the two-dimensional time-frequency energy distribution map and the differential spectrum matrix correspond in time. If the differential spectrum matrix corresponding to the two-dimensional time-frequency energy distribution map is calculated, a frequency instability vector containing values for four frequency bands is obtained, for example, [0.1, 0.2, 4.5, 0.2]. The instability vector shows that the signal in the third frequency band varies most dramatically, with an instability value of 4.5, while the other frequency bands are relatively stable. Normalizing the instability vector generates a frequency attention weight vector, which becomes [0.02, 0.04, 0.9, 0.04]. The time-frequency map consists of many rows and columns, with each row representing a specific frequency band. Multiplying the first row of data by a weight of 0.02, the second row by 0.04, the third row by 0.9, and the fourth row by 0.04 amplifies all features in the third frequency band, while features in other frequency bands are suppressed. This time-frequency graph with key frequency information is input into a two-dimensional convolutional neural network for feature extraction to obtain features that reflect the time-frequency characteristics of the signal.

[0030] In an optional embodiment, determining the fusion weight based on the fundamental amplitude and total harmonic distortion of the current data in the sliding window includes: Calculating the fundamental wave amplitude and total harmonic distortion of each current data in the sliding window; normalizing the fundamental wave amplitude and total harmonic distortion to obtain normalized fundamental wave amplitude and normalized total harmonic distortion; The normalized fundamental wave amplitude and the normalized total harmonic distortion are input into the weighting function to obtain the fusion weight.

[0031] Specifically, assume that the calculated average fundamental wave amplitude within the window is 10 A, which represents a medium load level. At the same time, the average total harmonic distortion is 30%, indicating significant harmonic pollution in the line. Map them into the normalized range from zero to one. Assume that the normal range of the fundamental wave amplitude is from 0 to 20 A, and the range of the total harmonic distortion is from 0 to 50%. Then the normalized fundamental wave amplitude is 0.5, and the normalized total harmonic distortion is 0.6. The normalized fundamental wave amplitude of 0.5 and the total harmonic distortion of 0.6 are input into a preset weighting function. After the weighting function performs operations, fusion weights are obtained. For example, 0.25 and 0.75 are output. This indicates that in the current environment of medium load and severe harmonic pollution, the reliability of the differential spectrum analysis channel is relatively low, while the results of the time-frequency analysis channel are more credible. In one embodiment, the weighting function is implemented using a piecewise function. First, establish rules. For example, one rule is: if the total harmonic distortion is very high, it indicates that the background noise of the line is large. At this time, the reliability of the differential spectrum characteristics decreases, and more trust should be placed in the characteristics of the time-frequency analysis. Then, determine the fusion weights according to the rules. In another embodiment, the weighting function is obtained through fitting functions such as multivariate polynomial regression and support vector regression using historical data.

[0032] The above embodiments are only used to illustrate the technical solutions of the present invention, not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention. In addition, any combination can be made among the various different embodiments of the embodiments of the present invention, as long as it does not violate the idea of the embodiments of the present invention, it should also be regarded as the content disclosed in the embodiments of the present invention.

Claims

1. A self-checking method for power distribution equipment based on an intelligent circuit breaker, characterized in that, Including: Obtain the current data of two consecutive full - cycle alternating currents monitored by the intelligent circuit breaker, perform full - phase Fourier transform on the current data of the two consecutive full - cycle alternating currents respectively to obtain two full - phase frequency spectra, and calculate a differential frequency spectrum matrix based on the two full - phase frequency spectra; calculate the spectral energy of the two full - phase frequency spectra within a preset high - frequency band respectively, take the current data of the full - cycle alternating current corresponding to the full - phase frequency spectrum with the maximum spectral energy as the target current data, perform continuous wavelet transform on the target current data to generate a time - frequency two - dimensional energy distribution map, and calculate a time - frequency information entropy value based on the time - frequency two - dimensional energy distribution map; Obtain a differential frequency spectrum matrix sequence and a time - frequency two - dimensional energy distribution map sequence according to a preset sliding window length and step size, obtain a time - frequency information entropy value sequence corresponding to the time - frequency two - dimensional energy distribution map sequence, obtain the channel attention of the differential frequency spectrum matrix sequence according to the time - frequency information entropy value sequence, and extract the features of the differential frequency spectrum matrix sequence by using the channel attention; Extract the features of the time - frequency two - dimensional energy distribution map sequence based on the stability of the frequency components in the differential frequency spectrum matrix; Determine a fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window, use the fusion weight to perform weighted fusion on the features of the differential frequency spectrum matrix sequence and the features of the time - frequency two - dimensional energy distribution map sequence, and obtain a self - inspection result based on the result of the weighted fusion.

2. The method according to claim 1, characterized in that, The obtaining the channel attention of the differential frequency spectrum matrix sequence according to the time - frequency information entropy value sequence includes: Input the time - frequency information entropy value sequence into a multi - layer perceptron, and the multi - layer perceptron includes at least one fully - connected layer and one activation layer; Use the multi - layer perceptron to perform a non - linear transformation on the time - frequency information entropy value sequence to obtain an attention weight vector with the same number as the differential frequency spectrum matrix sequence, and each value in the attention weight vector corresponds to a differential frequency spectrum matrix in the differential frequency spectrum matrix sequence.

3. The method according to claim 1, wherein The obtaining the channel attention of the differential frequency spectrum matrix sequence according to the time - frequency information entropy value sequence includes: Perform normalization processing on each entropy value in the time - frequency information entropy value sequence to obtain a normalized entropy value sequence; Multiply each matrix in the normalized entropy value sequence element - by - element with each matrix in the differential frequency spectrum matrix sequence to obtain a weighted differential frequency spectrum matrix sequence; Sum the rows and columns of each matrix in the weighted differential frequency spectrum matrix sequence to obtain a channel attention vector; Normalize the channel attention vector to obtain the channel attention.

4. The method according to claim 1, characterized in that, The extracting the features of the time - frequency two - dimensional energy distribution map sequence based on the stability of the frequency components in the differential frequency spectrum matrix includes: For each time - frequency two - dimensional energy distribution map in the time - frequency two - dimensional energy distribution map sequence, obtain the differential frequency spectrum matrix at the same position in the sliding window; Calculate the absolute value of each frequency component in the differential frequency spectrum matrix to obtain a frequency instability vector; Normalize the frequency instability vector to obtain a frequency attention weight vector; Multiply each row of the time - frequency two - dimensional energy distribution map by the corresponding weight value in the frequency attention weight vector to obtain an enhanced time - frequency two - dimensional energy distribution map; Input the enhanced sequence of time-frequency two-dimensional energy distribution diagrams into a two-dimensional convolutional neural network to obtain the features of the sequence of time-frequency two-dimensional energy distribution diagrams.

5. The method according to claim 1, characterized in that Determining the fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window includes: Calculating the fundamental wave amplitude and total harmonic distortion of each current data within the sliding window; performing normalization processing on the fundamental wave amplitude and total harmonic distortion to obtain the normalized fundamental wave amplitude and normalized total harmonic distortion; Inputting the normalized fundamental wave amplitude and normalized total harmonic distortion into a weighting function to obtain the fusion weight.

6. A self-checking system for power distribution equipment based on an intelligent circuit breaker, characterized in that, Including: A preprocessing unit for obtaining the current data of two consecutive full-cycle alternating currents monitored by an intelligent circuit breaker, respectively performing full-phase Fourier transform on the current data of the two consecutive full-cycle alternating currents to obtain two full-phase frequency spectra, calculating a differential frequency spectrum matrix based on the two full-phase frequency spectra; respectively calculating the spectral energy of the two full-phase frequency spectra within a preset high-frequency band, taking the current data of the full-cycle alternating current corresponding to the full-phase frequency spectrum with the maximum spectral energy as the target current data, performing continuous wavelet transform on the target current data to generate a time-frequency two-dimensional energy distribution diagram, and calculating the time-frequency information entropy value based on the time-frequency two-dimensional energy distribution diagram; A feature extraction unit for obtaining a sequence of differential frequency spectrum matrices and a sequence of time-frequency two-dimensional energy distribution diagrams according to a preset sliding window length and step size, obtaining a sequence of time-frequency information entropy values corresponding to the sequence of time-frequency two-dimensional energy distribution diagrams, obtaining the channel attention of the sequence of differential frequency spectrum matrices according to the sequence of time-frequency information entropy values, and extracting the features of the sequence of differential frequency spectrum matrices by using the channel attention; Extracting the features of the sequence of time-frequency two-dimensional energy distribution diagrams based on the stability of the frequency components in the differential frequency spectrum matrix; A self-check unit for determining the fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window, using the fusion weight to perform weighted fusion on the features of the sequence of differential frequency spectrum matrices and the features of the sequence of time-frequency two-dimensional energy distribution diagrams, and obtaining a self-check result based on the result of the weighted fusion.

7. The system according to claim 6, wherein The obtaining the channel attention of the sequence of differential frequency spectrum matrices according to the sequence of time-frequency information entropy values includes: Inputting the sequence of time-frequency information entropy values into a multi-layer perceptron, where the multi-layer perceptron includes at least one fully connected layer and one activation layer; Using the multi-layer perceptron to perform a non-linear transformation on the sequence of time-frequency information entropy values to obtain an attention weight vector with the same number as the sequence of differential frequency spectrum matrices, and each value in the attention weight vector corresponds to a differential frequency spectrum matrix in the sequence of differential frequency spectrum matrices.

8. The system according to claim 6, wherein The obtaining the channel attention of the sequence of differential frequency spectrum matrices according to the sequence of time-frequency information entropy values includes: Performing normalization processing on each entropy value in the sequence of time-frequency information entropy values to obtain a sequence of normalized entropy values; Multiplying each matrix in the sequence of normalized entropy values element-wise with each matrix in the sequence of differential frequency spectrum matrices to obtain a sequence of weighted differential frequency spectrum matrices; Performing row and column summation on each matrix in the sequence of weighted differential frequency spectrum matrices to obtain a channel attention vector; Normalizing the channel attention vector to obtain the channel attention.

9. The system according to claim 6, wherein Extracting the features of the time-frequency two-dimensional energy distribution map sequence based on the stability of the frequency components in the differential spectrum matrix includes: For each time-frequency two-dimensional energy distribution map in the time-frequency two-dimensional energy distribution map sequence, obtain the differential spectrum matrix at the same position in the sliding window; Calculate the absolute value of each frequency component in the differential spectrum matrix to obtain a frequency instability vector; Normalize the frequency instability vector to obtain a frequency attention weight vector; Multiply each row of the time-frequency two-dimensional energy distribution map by the corresponding weight value in the frequency attention weight vector to obtain an enhanced time-frequency two-dimensional energy distribution map; Input the enhanced time-frequency two-dimensional energy distribution map sequence into a two-dimensional convolutional neural network to obtain the features of the time-frequency two-dimensional energy distribution map sequence.

10. The system according to claim 6, characterized in that, Determining the fusion weight based on the fundamental wave amplitude and total harmonic distortion of the current data within the sliding window includes: Calculate the fundamental wave amplitude and total harmonic distortion of each current data within the sliding window; perform normalization processing on the fundamental wave amplitude and total harmonic distortion to obtain the normalized fundamental wave amplitude and normalized total harmonic distortion; Input the normalized fundamental wave amplitude and normalized total harmonic distortion into a weighting function to obtain the fusion weight.

Images