A method for detecting intermittent breakdown faults of electrical equipment

Abstract

The present invention discloses a method for detecting intermittent breakdown faults of electrical equipment, belonging to the technical field of power system fault detection, including: S1, collecting the current signal of the electrical equipment and adjusting the filter based on the current signal frequency and preset parameters; S2, inputting the current signal into the adjusted filter to obtain the decomposition mode; S3, iteratively updating the coefficients of the filter based on the decomposition mode and the iterative threshold to obtain the final decomposition mode; S4, performing safety analysis based on the final decomposition mode and the correlation coefficient criterion to obtain the safety analysis result; S5, determining the operating state of the corresponding electrical equipment based on the safety analysis result; The present invention significantly improves the accuracy of electrical equipment fault detection by introducing corresponding current characteristics in the process of electrical equipment fault detection.

Description

Technical Field

[0001] The present invention relates to the technical field of power system fault detection, and particularly to a method for detecting intermittent breakdown faults of electrical equipment. Background Art

[0002] In the early fault stage before an electrical equipment fails due to insulation deterioration, transient signals with intermittent breakdown characteristics often occur. By detecting the transient signals, faults of the electrical equipment can be discovered in advance, avoiding the chain reaction caused by electrical equipment faults. However, the transient signals with intermittent breakdown characteristics are extremely weak and rich in noise, making it extremely difficult to extract the transient signals, and thus the fault detection accuracy of the electrical equipment is low.

[0003] Chinese Patent, Publication No.: CN116401530A, Publication Date: July 7, 2023, discloses a bearing fault diagnosis method based on the optimization of feature mode decomposition by the WOA algorithm, including: acquiring the bearing fault signal to be diagnosed; adaptively searching for the optimal input parameter combination of the FMD algorithm by the WOA algorithm; decomposing the bearing fault signal by the FMD algorithm with the determined input parameters to obtain several modal components; calculating the kurtosis value of each modal component and selecting the modal component with the largest kurtosis value for envelope demodulation, and then extracting the corresponding fault feature information to achieve bearing fault diagnosis; and this invention does not consider the current characteristics, resulting in low fault detection accuracy of electrical equipment. Summary of the Invention

[0004] The object of the present invention is to address the problem that the prior art does not consider current characteristics, resulting in low fault detection accuracy of electrical equipment; and a method for detecting intermittent breakdown faults of electrical equipment is proposed. The filter is adjusted based on the current signal frequency and preset parameters, and the final decomposition mode is obtained by inversely iteratively trimming the filter with the decomposition mode obtained by processing the current signal with the filter. Then, safety analysis is performed according to the final decomposition mode and the correlation coefficient criterion to obtain a safety analysis result, and the operating state of the corresponding electrical equipment is determined based on the safety analysis result; by introducing corresponding current characteristics in the process of electrical equipment fault detection, the accuracy of electrical equipment fault detection is significantly improved.

[0005] In a first aspect, a technical solution provided in an embodiment of the present invention is a method for detecting intermittent breakdown faults of electrical equipment, including the following steps:

[0006] S1. Collect the current signal of the electrical equipment, and adjust the filter based on the current signal frequency and preset parameters;

[0007] S2. Input the current signal into the adjusted filter to obtain a decomposition mode;

[0008] S3. Iteratively update the coefficients of the filter based on the decomposition mode and the iterative threshold to obtain the final decomposition mode;

[0009] S4. Conduct a safety analysis based on the final decomposition mode and the correlation coefficient criterion to obtain the safety analysis result;

[0010] S5. Determine the operating state of the corresponding electrical equipment based on the safety analysis result.

[0011] In this solution, in order to introduce the current characteristics of the current signal during the fault detection of electrical equipment, the filter can evenly cover the current signal frequency band, and the number of filters is greater than the number of modes in the preset parameters, effectively avoiding the overlap or omission of the filter frequency bands; however, at this time, the filter may not be suitable for the specific working environment, which may lead to incorrect fault analysis. Therefore, it is necessary to perform iterative adjustment according to the decomposition mode. By updating the filter during the iterative adjustment process, the decomposition mode obtained by the filter processing the current signal can represent the fault state of the electrical equipment, and then the final decomposition mode can be obtained; different final decomposition modes may contain the same components. In order to eliminate the modal aliasing or redundancy, a safety analysis is conducted according to the final decomposition mode and the correlation coefficient criterion, and the final decomposition modes of the overlapping parts are removed to obtain an accurate safety analysis result. Finally, the operating state of the corresponding electrical equipment is determined based on the safety analysis result, significantly improving the accuracy of electrical equipment fault detection.

[0012] Preferably, in the S1, the adjustment formula for adjusting the filter based on the current signal frequency and the preset parameters is specifically:

[0013] , ;

[0014] , ;

[0015] , ;

[0016] In the formula, is the lower cut-off frequency, is the upper cut-off frequency, is the total number of filters, is the current signal frequency, is the window corresponding to the filter, is the time index of the window, is the length of the current signal, , is the filter length in the preset parameters.

[0017] In this solution, based on the adjustment formula, it can be seen that the current signal frequency is evenly divided into In the segment, corresponding upper cut-off frequency and lower cut-off frequency are set. At this time, the number of filters can be determined according to the number of segments of the current signal frequency. The frequency range of the filters is set by using the upper cut-off frequency and the lower cut-off frequency to achieve coverage of the current signal frequency band, effectively avoiding overlapping or omission of the filter frequency bands, and adjusting the window function corresponding to the filter by using the filter length and the number of modes in the preset parameters, so as to ensure that the interference information in the current signal, such as noise, can be accurately filtered out, and the decomposition state obtained by processing the current signal through the filter can accurately reflect the characteristics of the original current signal. Secondly, if the current signal is discrete data, the length of the current signal is the number of sampling data points corresponding to the sampling time and sampling frequency of the current signal. If the current signal is continuous data, the length of the current signal is the duration of the current signal. The current signal at least includes power frequency current signal, high-frequency fault characteristics, and noise. The preset parameters at least include the number of modes, filter length, and iteration threshold set based on the electrical equipment.

[0018] Preferably, in the S2, the specific decomposition formula for obtaining the decomposition mode by inputting the current signal into the adjusted filter is as follows:

[0019] ;

[0020] In the formula, is the time index of the th iteration of the th decomposition mode, is the th iteration of the th decomposition mode, is the current signal, is the th iteration of the th decomposition mode corresponding to the filter coefficient.

[0021] In this solution, the time index is actually the output time corresponding to the decomposition mode.

[0022] Preferably, in the S3, the specific process of iteratively updating the filter coefficient based on the decomposition mode and the iteration threshold to obtain the final decomposition mode is as follows:

[0023] S31. Estimate the input period of the decomposition mode based on the autocorrelation function of the decomposition mode, and calculate the correlation kurtosis according to the input period and the decomposition mode;

[0024] S32. Update the filter coefficients based on the weighted conversion of the relevant kurtosis, and determine whether to end the update based on the update times of the filter coefficients and the iteration threshold. If the update times are less than the iteration threshold, input the current signal into the updated filter to obtain the iterative decomposition mode, update the decomposition mode based on the iterative decomposition mode, and execute S31. If the update times are greater than or equal to the iteration threshold, mark the decomposition mode with the maximum relevant kurtosis as the final decomposition mode.

[0025] In this solution, since a corresponding fault component will be generated when the electrical equipment fails, the decomposition mode corresponding to the fault component has the characteristic of the maximum relevant kurtosis. By updating the filter coefficients based on the weighted conversion of the relevant kurtosis during the iterative adjustment process, it is ensured that the update of the filter coefficients is based on the relevant kurtosis as large as possible, so that the decomposition mode obtained by the filter processing the current signal can be closest to the fault component to the greatest extent. Furthermore, the final decomposition mode can represent the fault state of the electrical equipment, and the coefficients of the filter can also fit the working environment. The relevant kurtosis can be used as a characteristic parameter to measure the outlier degree of outlier data and can also be used to reflect the periodicity and impact of the signal.

[0026] Preferably, in S31, the autocorrelation function is specifically:

[0027] ;

[0028] In the formula, is the autocorrelation function value at , is the time index of the corresponding decomposition mode, is the length of the current signal corresponding to the decomposition mode, is the current signal at is the current signal at

[0029] In this solution, when the autocorrelation function value is the maximum value, the corresponding time point is the input period. Therefore, the time point corresponding to the maximum value after the first zero crossing is used as the input period.

[0030] Preferably, in S31, the relevant kurtosis formula corresponding to the relevant kurtosis is specifically:

[0031] ;

[0032] In the formula, is the relevant kurtosis corresponding to the decomposition mode , is the time index of the corresponding decomposition mode, is the length of the current signal corresponding to the decomposition mode, is the input period.

[0033] Preferably, in the step S32, the update formula for updating the filter coefficients based on the correlation kurtosis is as follows:

[0034] ;

[0035] ;

[0036] In the formula, is the correlation kurtosis, is the filter coefficient matrix, is the current signal matrix, is the weighting matrix, represents the conjugate transpose operation, , is the updated coefficient, , is the eigenvalue.

[0037] Preferably, in the step S4, the specific process of performing safety analysis according to the final decomposition mode and the correlation coefficient criterion to obtain the safety analysis result is as follows:

[0038] S41. Input the final decomposition mode into the correlation coefficient formula to calculate the correlation coefficient, and sort the correlation coefficients by size to obtain a correlation coefficient sequence;

[0039] S42. Extract the correlation kurtosis of the final decomposition mode corresponding to the largest correlation coefficient from the correlation coefficient sequence to obtain the largest correlation kurtosis, and extract the correlation kurtosis of the final decomposition mode corresponding to the second largest correlation coefficient from the correlation coefficient sequence to obtain the second largest correlation kurtosis;

[0040] S43. Compare the largest correlation kurtosis with the second largest correlation kurtosis. If the largest correlation kurtosis is less than the second largest correlation kurtosis, then eliminate the final decomposition mode corresponding to the largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the largest correlation kurtosis is greater than the second largest correlation kurtosis, then eliminate the final decomposition mode corresponding to the second largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the largest correlation kurtosis is equal to the second largest correlation kurtosis, then establish a safety label and end the analysis;

[0041] S44. Determine whether to end the analysis based on the total number of filters and the number of modes in the preset parameters. If the total number of filters is greater than the number of modes, then execute S41. If the total number of filters is less than or equal to the number of modes, then mark the corresponding final decomposition mode as a fault component and end the analysis;

[0042] S45. After the analysis is completed, organize the safety label and the fault components to obtain the safety analysis result.

[0043] In this solution, it is necessary to calculate the correlation coefficients of different modes. When the correlation coefficient is too large, the similarity of the corresponding two modes is high, and there is a situation of mode aliasing or redundancy. At this time, compare the correlation kurtosis of the two modes. Based on the characteristic that the correlation kurtosis corresponding to the fault components appears to be the largest when the electrical equipment fails, eliminate the final decomposition mode with a smaller correlation kurtosis, and at the same time reduce the total number of filters by one. And when the total number of filters is greater than the number of modes in the preset parameters, loop to calculate the correlation coefficient and delete the final decomposition mode until the total number of filters is less than or equal to the number of modes and then stop. Secondly, when the correlation kurtosis corresponding to the two modes with the largest correlation coefficient is equal, it proves that the two modes are exactly the same, and when the correlation kurtosis does not appear to be the largest, it proves that the corresponding electrical equipment has not failed. At this time, it is necessary to establish a safety label and end the analysis. Otherwise, if a larger correlation kurtosis can always be compared until there is still a largest correlation kurtosis at the end of the analysis, it proves that the corresponding electrical equipment has failed, and mark the corresponding final decomposition mode as a fault component.

[0044] Preferably, in S41, the specific formula for the correlation coefficient is:

[0045] ;

[0046] In the formula, is the correlation coefficient between the final decomposition mode and the final decomposition mode , is the final decomposition mode when , , is the length of the corresponding current signal, is the time period the average value of all the final decomposition modes within, is when the final decomposition mode , is the time period the average value of all the final decomposition modes within.

[0047] Preferably, in S5, the specific process of determining the operating state of the corresponding electrical equipment based on the safety analysis result is as follows:

[0048] Count the number of safety labels in the safety analysis result and count the number of fault components in the safety analysis result;

[0049] Judge the fault state of the corresponding electrical equipment based on the number of safety tags and the number of fault components. If the number of safety tags is greater than the safety threshold and the number of fault components is less than the fault threshold, it is determined that the corresponding electrical equipment is in a normal state. If the number of safety tags is less than or equal to the safety threshold and the number of fault components is greater than or equal to the fault threshold, it is determined that the corresponding electrical equipment has a fault.

[0050] Advantages of the present invention:

[0051] (1) In this application, the frequency of the current signal is evenly divided into segments, and the corresponding upper cut-off frequency and lower cut-off frequency are set. At this time, the number of filters can be determined according to the number of segments of the current signal frequency. The frequency range of the filter is set by using the upper cut-off frequency and the lower cut-off frequency to achieve coverage of the current signal frequency band, effectively avoiding overlapping or omission of the filter frequency band, and adjusting the window function corresponding to the filter by using the filter length and the number of modes in the preset parameters, so as to ensure that the interference information in the current signal, such as noise, can be accurately filtered out, and the decomposition state obtained by processing the current signal through the filter can accurately reflect the characteristics of the original current signal;

[0052] (2) In this application, the filter is iteratively adjusted according to the kurtosis of the relevant decomposition modes, and based on the characteristic that the kurtosis corresponding to the fault component when the electrical equipment fails is the largest, the filter coefficients are updated by weighted conversion based on the kurtosis during the iterative adjustment process, ensuring that the update of the filter coefficients is based on the kurtosis as large as possible, so that the decomposition mode obtained by the filter processing the current signal can be closest to the fault component to the greatest extent, and the coefficients of the filter can also fit the working environment, and then the final decomposition mode corresponding to the fault component is obtained, effectively improving the comprehensiveness of the extraction of the corresponding fault component when the electrical equipment fails;

[0053] (3) In this application, safety analysis is performed according to the final decomposition mode and the correlation coefficient criterion to obtain a safety analysis result. During the safety analysis process, the kurtosis of the two modes with high correlation coefficients is compared. If the kurtosis is equal, based on the characteristic that the kurtosis corresponding to the fault component when the electrical equipment fails is the largest, the absence of the largest kurtosis proves that the electrical equipment has not failed, and a safety tag is established. If the kurtosis is not equal, the final decomposition mode with a small kurtosis is removed, and the number of modes is adjusted until the fault component corresponding to the electrical equipment is extracted. Finally, the occurrence times of the safety tags and the fault components in the safety analysis result are counted to accurately judge the operating state of the corresponding electrical equipment, significantly improving the accuracy of electrical equipment fault detection. Description of the Drawings

[0054] Other features, objects, and advantages of the present invention will become more apparent from the detailed description of non - restrictive embodiments read in conjunction with the following drawings. The drawings are only for the purpose of showing the preferred embodiments and are not to be considered as limiting the present invention. Also, throughout the drawings, the same reference numerals are used to denote the same components.

[0055] Figure 1 It is a schematic flow chart of a method for detecting intermittent breakdown faults of an electrical device;

[0056] Figure 2 It is a schematic waveform diagram of a current signal. Detailed Embodiments

[0057] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only the best embodiments of the present invention, which are only used to explain the present invention and do not limit the protection scope of the present invention. All other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0058] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operations are completed, but it can also have additional steps not included in the drawings; the process can correspond to a method, function, procedure, subroutine, sub - program, etc.

[0059] Embodiment 1:

[0060] As Figure 1 shown, this embodiment provides a method for detecting intermittent breakdown faults of an electrical device, including the following steps:

[0061] S1. Collect the current signal of the electrical device and adjust the filter based on the current signal frequency and preset parameters;

[0062] The corresponding adjustment formula is specifically:

[0063] ;

[0064] ;

[0065] ;

[0066] In the formula, is the lower cut-off frequency, is the upper cut-off frequency, is the total number of filters, is the current signal frequency, is the window corresponding to the filter, is the time index of the window, is the length of the current signal, , is the filter length in the preset parameters.

[0067] In this embodiment, as Figure 2 shown, the length of the current signal is 1000, which fluctuates in the range of 2A to -2A, and it can be clearly seen from the waveform of the current signal that pulse-like fault characteristics will appear in the second half of each half cycle, and there is also a large amount of noise in the current signal; and parameters for adjusting the filter are preset. Specifically, the number of modes is 1, the filter length is 40, and the iteration threshold is set to 20 at the same time; then the frequency of the current signal is evenly divided into 5 segments by using the adjustment formula, and a FIR filter is set for each segment, and a total of 5 FIR filters are set, and the 5 FIR filters together form a filter bank. At this time, each filter also corresponds to an initial coefficient. Secondly, the window function is selected as the Hamming window, which can be expressed as: , .

[0068] S2. Input the current signal into the adjusted filter to obtain the decomposed mode;

[0069] The corresponding decomposition formula is specifically:

[0070] ;

[0071] In the formula, is the time index of the th iteration of the th decomposed mode, is the th iteration of the th decomposed mode, is the current signal, is the th iteration of the th decomposed mode corresponding to the filter coefficient.

[0072] S3. Iteratively update the filter coefficients based on the decomposed mode and the iteration threshold to obtain the final decomposed mode;

[0073] S31. Estimate the input period of the decomposed mode based on the autocorrelation function of the decomposed mode, and calculate the correlation kurtosis according to the input period and the decomposed mode;

[0074] The autocorrelation function is specifically as follows:

[0075] ;

[0076] In the formula, is the autocorrelation function value at is the time index of the corresponding decomposition mode, is the length of the current signal corresponding to the decomposition mode, is the current signal at is the current signal at;

[0077] The correlation kurtosis formula corresponding to the correlation kurtosis is specifically as follows:

[0078] ;

[0079] In the formula, is the correlation kurtosis corresponding to the decomposition mode , is the time index of the corresponding decomposition mode, is the length of the current signal corresponding to the decomposition mode, is the input period;

[0080] S32. Update the filter coefficients based on the correlation kurtosis for weighted conversion. Judge whether to end the update based on the update times of the filter coefficients and the iteration threshold. If the update times are less than the iteration threshold, input the current signal into the updated filter to obtain the iterative decomposition mode, update the decomposition mode based on the iterative decomposition mode and execute S31. If the update times are greater than or equal to the iteration threshold, mark the decomposition mode with the maximum correlation kurtosis as the final decomposition mode;

[0081] The update formula corresponding to updating the filter coefficients based on the correlation kurtosis for weighted conversion is:

[0082] ;

[0083] ;

[0084] In the formula, is the correlation kurtosis, is the filter coefficient matrix, is the current signal matrix, is the weighted matrix, represents the conjugate transpose operation, , is the updated coefficient, , is the eigenvalue.

[0085] In this embodiment, the basis for updating the filter coefficients is the relevant kurtosis of the decomposition mode. The relevant kurtosis is a characteristic parameter for measuring the degree of outliers of outlier data, which can reflect the periodicity and impact of the current signal. Generally, the relevant kurtosis corresponding to the fault component generated when an electrical device fails is the largest. For the convenience of calculation, the decomposition formula and the relevant kurtosis formula can be converted into matrix forms. The matrix form of the decomposition formula is specifically as follows:

[0086] ;

[0087] In the formula, is the th iteration of the th decomposition mode, is the current signal, is the th iteration of the th decomposition mode corresponding to the filter coefficient;

[0088] The matrix form of the relevant kurtosis formula is specifically as follows:

[0089] ;

[0090] In the formula, is the relevant kurtosis corresponding to the decomposition mode , is the th iteration of the th decomposition mode, represents the conjugate transpose operation, is the weighted correlation matrix;

[0091] Specifically, the decomposition mode, coefficient, and current signal can be expressed as:

[0092] ;

[0093] ;

[0094] ;

[0095] The weighted correlation matrix can be specifically expressed as:

[0096] ;

[0097] The simultaneous equations relate the relevant kurtosis to the filter coefficients to obtain an update formula. When solving the update formula, in order to make the decomposition mode obtained by the filter processing the current signal close to the fault component, or even directly filter out the fault component, the objective function, that is, the relevant kurtosis corresponding to the decomposition mode should be as large as possible. Therefore, in the iterative update process, the filter coefficients corresponding to the maximum relevant kurtosis are solved in each iteration, which is equivalent to solving the eigenvector related to the maximum eigenvalue in the generalized eigenvalue problem. Therefore, the solution method of the update formula is the same as that of solving the generalized eigenvalue problem. After the solution is completed, the filter coefficients, the corresponding decomposition mode, the weighting matrix, etc. are changed to achieve the gradual update of the adjoint iterative process until the update times are greater than or equal to the iteration threshold, and the decomposition mode with the largest relevant kurtosis is marked as the final decomposition mode.

[0098] S4. Perform a safety analysis based on the final decomposition mode and the correlation coefficient criterion to obtain a safety analysis result;

[0099] S41. Input the final decomposition mode into the correlation coefficient formula to calculate the correlation coefficient, and sort the correlation coefficients to obtain a correlation coefficient sequence;

[0100] The specific correlation coefficient formula is:

[0101] ;

[0102] In the formula, is the final decomposition mode and the final decomposition mode The correlation coefficient between them, is The final decomposition mode at , is the length of the corresponding current signal, is the time period The average value of all final decomposition modes within, is The final decomposition mode at , is the time period The average value of all final decomposition modes within;

[0103] S42. Extract the relevant kurtosis of the final decomposition mode corresponding to the first largest correlation coefficient from the correlation coefficient sequence to obtain the first largest relevant kurtosis, and extract the relevant kurtosis of the final decomposition mode corresponding to the second largest correlation coefficient from the correlation coefficient sequence to obtain the second largest relevant kurtosis;

[0104] S43. Compare the first largest correlation kurtosis with the second largest correlation kurtosis. If the first largest correlation kurtosis is less than the second largest correlation kurtosis, then eliminate the final decomposition mode corresponding to the first largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the first largest correlation kurtosis is greater than the second largest correlation kurtosis, then eliminate the final decomposition mode corresponding to the second largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the first largest correlation kurtosis is equal to the second largest correlation kurtosis, then establish a security label and end the analysis;

[0105] S44. Determine whether to end the analysis based on the total number of filters and the number of modes in the preset parameters. If the total number of filters is greater than the number of modes, then execute S41. If the total number of filters is less than or equal to the number of modes, then mark the corresponding final decomposition mode as a fault component and end the analysis;

[0106] S45. After the analysis ends, organize the security label and the fault component to obtain a security analysis result.

[0107] S5. Determine the operating state of the corresponding electrical equipment based on the security analysis result;

[0108] Count the number of security labels in the security analysis result, and count the number of fault components in the security analysis result;

[0109] Judge the fault state of the corresponding electrical equipment based on the number of security labels and the number of fault components. If the number of security labels is greater than the security threshold and the number of fault components is less than the fault threshold, then determine that the corresponding electrical equipment is in a normal state. If the number of security labels is less than or equal to the security threshold and the number of fault components is greater than or equal to the fault threshold, then determine that the corresponding electrical equipment has a fault.

[0110] This embodiment at least has the following substantial effects:

[0111] (1) In this application, the frequency of the current signal is evenly divided into segments, and the corresponding upper cut-off frequency and lower cut-off frequency are set. At this time, the number of filters can be determined according to the number of segments of the current signal frequency. The frequency range of the filter is set by using the upper cut-off frequency and the lower cut-off frequency to achieve coverage of the current signal frequency band, effectively avoiding overlap or omission of the filter frequency band, and adjusting the window function corresponding to the filter by using the filter length and the number of modes in the preset parameters, so as to ensure that the interference information in the current signal, such as noise, can be accurately filtered out, and the decomposition state obtained by processing the current signal through the filter can accurately reflect the characteristics of the original current signal;

[0112] (2) This application iteratively adjusts the filter according to the kurtosis of the decomposed modes, and based on the characteristic that the kurtosis corresponding to the fault components that appear during the electrical equipment failure is the largest. By performing weighted transformation to update the filter coefficients based on the kurtosis during the iterative adjustment process, it ensures that the update of the filter coefficients is based on the kurtosis being as large as possible, enabling the decomposed modes obtained by the filter processing the current signal to be as close as possible to the fault components to the greatest extent. It can also make the filter coefficients fit the working environment, thereby obtaining the final decomposed modes corresponding to the fault components, effectively improving the comprehensiveness of extracting the corresponding fault components during the electrical equipment failure.

[0113] (3) This application performs safety analysis based on the final decomposed modes and the correlation coefficient criterion to obtain the safety analysis results. During the safety analysis process, the kurtosis of the two modes with high correlation coefficients is compared. If the kurtosis is equal, based on the characteristic that the kurtosis corresponding to the fault components that appear during the electrical equipment failure is the largest, the absence of the largest kurtosis proves that the electrical equipment has not failed, and a safety label is established. If the kurtosis is not equal, the final decomposed mode with a small kurtosis is removed, and the number of modes is adjusted until the fault components corresponding to the electrical equipment are extracted. Finally, by counting the occurrence times of the safety labels and the fault components in the safety analysis results, the operating state of the corresponding electrical equipment is accurately judged, significantly improving the accuracy of electrical equipment fault detection.

[0114] The above specific implementation manners are the preferred implementation manners of the present invention, and do not limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to this specific implementation manner. All equivalent changes made according to the shape, structure, and method of the present invention are within the protection scope of the present invention.

Claims

1. A method for detecting intermittent breakdown faults of an electrical device, characterized in that, It includes the following steps: S1. Collect the current signal of the electrical equipment and adjust the filter based on the current signal frequency and preset parameters; The adjustment formula for adjusting the filter corresponding to the current signal frequency and preset parameters is specifically: where f lk is the lower cut-off frequency, f uk is the upper cut-off frequency, K is the total number of filters, f s is the current signal frequency, w(n1) is the window corresponding to the filter, n1 is the time index of the window, N is the length of the current signal, N0 = L - 1, and L is the filter length in the preset parameters; S2. Input the current signal into the adjusted filter to obtain the decomposition mode; S3. Iteratively update the coefficients of the filter based on the decomposition mode and the iteration threshold to obtain the final decomposition mode; S31. Estimate the input period of the decomposition mode based on the autocorrelation function of the decomposition mode, and calculate the correlation kurtosis according to the input period and the decomposition mode; S32. Perform weighted conversion based on the correlation kurtosis to update the filter coefficients, and determine whether to end the update based on the update times of the filter coefficients and the iteration threshold. If the update times are less than the iteration threshold, input the current signal into the updated filter to obtain the iterative decomposition mode, update the decomposition mode based on the iterative decomposition mode and execute S31. If the update times are greater than or equal to the iteration threshold, mark the decomposition mode with the maximum correlation kurtosis as the final decomposition mode; S4. Perform safety analysis based on the final decomposition mode and the correlation coefficient criterion to obtain the safety analysis result; S41. Input the final decomposition mode into the correlation coefficient formula to calculate the correlation coefficient, and sort the correlation coefficients by size to obtain the correlation coefficient sequence; S42. Extract the correlation kurtosis corresponding to the first largest correlation coefficient of the final decomposition mode based on the correlation coefficient sequence to obtain the first largest correlation kurtosis, and extract the correlation kurtosis corresponding to the second largest correlation coefficient of the final decomposition mode based on the correlation coefficient sequence to obtain the second largest correlation kurtosis; S43. Compare the sizes of the first largest correlation kurtosis and the second largest correlation kurtosis. If the first largest correlation kurtosis is less than the second largest correlation kurtosis, eliminate the final decomposition mode corresponding to the first largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the first largest correlation kurtosis is greater than the second largest correlation kurtosis, eliminate the final decomposition mode corresponding to the second largest correlation kurtosis, subtract one from the total number of filters corresponding to the final decomposition mode, and execute S44. If the first largest correlation kurtosis is equal to the second largest correlation kurtosis, establish a safety label and end the analysis; S44. Determine whether to end the analysis based on the total number of filters and the number of modes in the preset parameters. If the total number of filters is greater than the number of modes, execute S41. If the total number of filters is less than or equal to the number of modes, mark the corresponding final decomposition mode as a fault component and end the analysis; S45. After the analysis is completed, organize the safety label and the fault component to obtain the safety analysis result; S5. Determine the operating state of the corresponding electrical equipment based on the safety analysis result.

2. The intermittent breakdown fault detection method for an electrical device according to claim 1, wherein In S2, the decomposition formula for inputting the current signal into the adjusted filter to obtain the decomposition mode is specifically: where n2 is the time index of the k-th decomposition mode in the i-th iteration, is the k-th decomposition mode in the i-th iteration, I(n2) is the current signal, are the coefficients of the filter corresponding to the k-th decomposition mode at the i-th iteration.

3. A method for detecting intermittent breakdown faults of an electrical device according to claim 1, characterized in that, In S31, the autocorrelation function is specifically: where R x (τ) is the autocorrelation function value at τ, n2 is the time index of the corresponding decomposition mode, N is the length of the current signal corresponding to the decomposition mode, I(n2) is the current signal at n2, and I(n2+τ) is the current signal at n2+τ.

4. A method for detecting intermittent breakdown faults of an electrical device according to claim 1, characterized in that, In S31, the correlation kurtosis formula corresponding to the correlation kurtosis is specifically: In the formula, is the decomposition mode The corresponding relative kurtosis, n2 is the time index of the corresponding decomposition mode, N is the length of the current signal corresponding to the decomposition mode, and T s is the input period.

5. A method for detecting intermittent breakdown faults of an electrical device according to claim 1, characterized in that, In S32, the update formula for updating the filter coefficients by weighted conversion based on the correlation kurtosis is: where p(·) is the relevant kurtosis, is the filter coefficient matrix, I is the current signal matrix, and W i is the weighting matrix, H represents the conjugate transpose operation, is the updated coefficient, Q = I H × I, and λ is the eigenvalue.

6. A method for detecting intermittent breakdown faults of an electrical device according to claim 1, characterized in that, In S41, the correlation coefficient formula is specifically: where r pq is the correlation coefficient between the final decomposition mode p and the final decomposition mode q, u p (n2) is the final decomposition mode p at n2, N is the length of the corresponding current signal, is the average value of all the final decomposition modes p within the time period N, u q (n2) is the final decomposition mode q at n2, is the average value of all the final decomposition modes q within the time period N.

7. A method for detecting intermittent breakdown faults of an electrical device according to claim 1, characterized in that, In S5, the specific process of determining the operating state of the corresponding electrical equipment based on the safety analysis result is: Count the number of safety labels in the statistical safety analysis results and count the number of fault components in the statistical safety analysis results; Judge the fault status of the corresponding electrical equipment based on the number of safety labels and the number of fault components. If the number of safety labels is greater than the safety threshold and the number of fault components is less than the fault threshold, it is determined that the corresponding electrical equipment is in a normal state. If the number of safety labels is less than or equal to the safety threshold and the number of fault components is greater than or equal to the fault threshold, it is determined that the corresponding electrical equipment has a fault.

Images