Low-current arc fault detection method based on time-frequency domain disturbance and related equipment

Abstract

The invention discloses a low-current arc fault detection method based on time-frequency domain disturbance and related equipment, and the method comprises the steps: recognizing a time point where disturbance changes occur through an improved CUSUM algorithm, positioning a current data standard deviation, and then achieving the arc fault detection according to the changes of even harmonics before and after the time point. The circuit breaker solves the problems that a traditional circuit breaker can only detect short-circuit faults and overload faults by means of a line current rigid threshold value, and small-current arc faults are often missed and falsely detected due to small current change amplitude.

Description

Technical Field

[0001] The present invention relates to the technical field of small current arc fault detection, and in particular to a small current arc fault detection method based on time-frequency domain disturbance and related equipment. Background Art

[0002] In the current AC power distribution system, the long-term distribution lines show obvious signs of aging, and the insulation degradation and poor contact of the lines can easily lead to arc faults. In the early stage of an arc fault, due to the large impedance of the fault circuit, the change range of the fault current is smaller than that of the normal current. When there are other working loads in the same line, the fault current will be mixed with other load currents.

[0003] Existing technologies mainly rely on the physical characteristics and current characteristics of arcs for detection. Physical characteristics include radiation, arc light, and arc sound, but these methods are limited by the sensor installation location and are often used for fault arc detection of specific equipment. Current characteristic detection identifies fault arcs by measuring the main circuit current, extracting the time domain, frequency domain, and time-frequency domain characteristics of arc current using tools such as Fourier transform and wavelet transform, and setting empirical thresholds or classification algorithms for fault arc detection.

[0004] Since the fault circuit impedance is large at the beginning of an arc fault, the fault current changes less than the normal current and may be mixed with other workload currents. As a result, traditional circuit breakers relying on the line current rigid threshold have difficulty detecting small current arc faults, and missed detections and false detections often occur.

[0005] Therefore, at present, traditional circuit breakers can only detect short-circuit faults and overload faults by relying on the rigid threshold of line current. Small current arc faults are often missed or misdetected due to the small current change amplitude. Summary of the invention

[0006] Based on the problems raised by the above background technology, the purpose of the present invention is to provide a small current arc fault detection method and related equipment based on time-frequency domain disturbance, which realizes the rapid positioning of current disturbance by improving the CUSUM algorithm, and realizes the detection of small current arc fault by combining the time-frequency domain feature quantity, thereby solving the problem that the current traditional circuit breaker can only realize the detection of short circuit fault and overload fault by relying on the rigid threshold of line current, and small current arc fault is often missed or misdetected due to the small current change amplitude.

[0007] The present invention is achieved through the following technical solutions:

[0008] The first aspect of the present invention provides a small current arc fault detection method based on time-frequency domain disturbance, comprising the following steps:

[0009] Acquire current data, and use an improved CUSUM algorithm to perform current disturbance judgment on the current data;

[0010] Based on the current disturbance judgment result, locating the current disturbance starting point through the standard deviation and kurtosis of the current data;

[0011] The current data of the power frequency cycles before and after the starting point of the current disturbance are intercepted, and the arc fault detection is performed on the current data of the power frequency cycles before and after using the even harmonic technology to obtain the detection result.

[0012] In the above technical solution, the CUSUM algorithm is a statistical method for detecting change points in a data sequence, and the algorithm detects the degree of deviation of the data by calculating the accumulation of the data sequence. The small current arc fault detection method proposed in the present invention improves the characteristics of the CUSUM algorithm, combines it with a sliding window, and applies it to the small current arc fault detection method proposed in the present invention, and changes in the data sequence in the current data, thereby realizing the identification of the current disturbance.

[0013] The improved CUSUM algorithm can be used to determine whether the detected current data is disturbed. However, since the improved CUSUM algorithm needs to continuously accumulate the obtained offset values, the judgment result of the current disturbance obtained has obvious time lag. Therefore, this method proposes to locate the starting point of the disturbance change based on the current disturbance judgment result by the standard deviation and kurtosis of the current data, so as to realize the interception of the current data.

[0014] Then, the current data of the power frequency cycle before and after the starting point of the disturbance change is intercepted, and then the current data of the power frequency cycle before and after is used to detect arc faults, thereby achieving accurate current disturbance judgment. This method can realize fault detection for small current arcs, thereby filling the algorithm defects of the current traditional circuit breaker based on rigid threshold detection.

[0015] In an optional embodiment, using an improved CUSUM algorithm to perform current disturbance judgment on the current data includes the following steps:

[0016] Setting connected double-slide rail windows W1 and W2, and preprocessing the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window;

[0017] Obtain the distribution expectation threshold of the front window W1 and determine the offset threshold;

[0018] Based on the expected distribution value and the offset threshold, sequentially calculating the offset accumulation value of each data point of the preprocessed current data in the rear window W2, and iterating the offset accumulation value;

[0019] An offset limit is set, and the offset accumulation value after iteration is compared with the offset limit. If the comparison condition is met, it is determined that a current disturbance occurs in the rear window W2.

[0020] In an optional embodiment, preprocessing the current data intercepted by the double-slide window includes the following steps:

[0021] Calculate the effective value of the current data captured by the front window W1 and the rear window W2;

[0022] The average value of the effective value of the current data captured by the front window W1 is used as the reference value, and the effective values ​​of the current data captured by the front window W1 and the rear window W2 are calculated respectively to obtain the current normalization value.

[0023] In an optional embodiment, after determining that a current disturbance occurs in the rear window W2, the method further includes: the double-track window slides and shifts Y positions, wherein M<Y<2M, and M is the window length of the front window W1 and the rear window W2;

[0024] After the double-slide window slides and shifts Y positions, the offset accumulation value of the rear window W2 is initialized.

[0025] In an optional embodiment, based on the current disturbance judgment result, locating the current disturbance starting point by the standard deviation and kurtosis of the current data comprises the following steps:

[0026] All current normalized values ​​in the front window W1 and the rear window W2 are constructed into a current normalized value sequence;

[0027] Overlapping and sorting the current normalized value sequence to construct an effective value matrix, and calculating the standard deviation and kurtosis in each part of the current normalized value sequence through the effective value matrix to obtain a standard deviation sequence and a kurtosis sequence;

[0028] The data point with the largest standard deviation is selected from the standard deviation sequence as the center point, and k kurtosis values ​​are extracted from the kurtosis sequence with the center point as the center. If the kurtosis value corresponding to the center point is the local minimum of the k kurtosis values, then the center point is the starting point of the current disturbance.

[0029] In an optional embodiment, arc fault detection is performed on current data of previous and next power frequency cycles using even harmonic technology, including the following steps:

[0030] The current data of the power frequency cycles before and after the front window W1 and the rear window W2 are processed by short-time Fourier transform to obtain the front even harmonic sum sequence and the rear even harmonic sum sequence;

[0031] Calculating an even harmonic sum ratio of the sum of the preceding even harmonic sum sequence and the sum of the following even harmonic sum sequence;

[0032] A detection threshold is set, and if the even harmonic sum ratio is greater than the detection threshold, an arc fault exists.

[0033] A second aspect of the present invention provides a low current arc fault detection system based on time-frequency domain disturbance, comprising:

[0034] A current disturbance module is used to obtain current data and use an improved CUSUM algorithm to perform current disturbance judgment on the current data;

[0035] A starting point module, used for locating the starting point of the current disturbance through the standard deviation and kurtosis of the current data based on the current disturbance judgment result;

[0036] The fault detection module is used to intercept the current data of the power frequency cycle before and after the starting point of the current disturbance, and use the even harmonic technology to perform arc fault detection on the current data of the power frequency cycle before and after to obtain the detection result.

[0037] In an optional embodiment, the current disturbance module includes:

[0038] A preprocessing unit, used to set the connected double-slide rail windows W1 and W2, and preprocess the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window;

[0039] A front window unit, used to obtain a distribution expectation threshold of the front window W1 and determine an offset threshold;

[0040] A rear window unit, used to sequentially calculate the offset accumulation value of each data point of the preprocessed current data in the rear window W2 based on the distribution expected value and the offset threshold, and iterate the offset accumulation value;

[0041] The comparison unit is used to set an offset limit, compare the offset accumulation value after iteration with the offset limit, and if the comparison condition is met, determine that a current disturbance occurs in the rear window W2.

[0042] A third aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a small current arc fault detection method based on time-frequency domain disturbance is implemented.

[0043] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a small current arc fault detection method based on time-frequency domain disturbance.

[0044] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0045] The present invention can realize fault detection for small current arcs, thereby filling the algorithm defects of the current traditional circuit breaker based on rigid threshold detection; and realize accurate current disturbance position judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0047] Figure 1 This is an example diagram of sliding window displacement provided in Example 1 of the present invention;

[0048] Figure 2 A current-standard deviation-kurtosis sequence diagram provided in Example 1 of the present invention;

[0049] Figure 3 (a) and Figure 3 (b) is a schematic diagram of an application scenario of a small current arc fault detection method based on time-frequency domain disturbance provided in Example 1 of the present invention;

[0050] Figure 4 A schematic diagram of the structure of an electronic device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0051] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0052] Example 1

[0053] Embodiment 1 of the present invention provides a small current arc fault detection method based on time-frequency domain disturbance, and the fault detection method comprises the following steps:

[0054] Acquire current data, and use an improved CUSUM algorithm to perform current disturbance judgment on the current data;

[0055] Based on the current disturbance judgment result, locating the current disturbance starting point through the standard deviation and kurtosis of the current data;

[0056] The current data of the power frequency cycles before and after the starting point of the current disturbance are intercepted, and the arc fault detection is performed on the current data of the power frequency cycles before and after using the even harmonic technology to obtain the detection result.

[0057] It should be noted that the CUSUM algorithm is a statistical method for detecting change points in a data sequence. The algorithm detects the degree of deviation of the data by calculating the accumulation of the data sequence. The small current arc fault detection method proposed in the present invention uses the characteristics of the CUSUM algorithm to improve it, combines it with a sliding window, and applies it to the small current arc fault detection method proposed in the present invention, and changes in the data sequence in the current data, thereby realizing the identification of current disturbances.

[0058] The improved CUSUM algorithm can be used to determine whether the detected current data is disturbed. However, since the improved CUSUM algorithm needs to continuously accumulate the obtained offset values, the judgment result of the current disturbance obtained has obvious time lag. Therefore, this method proposes to locate the starting point of the disturbance change based on the current disturbance judgment result by the standard deviation and kurtosis of the current data, so as to realize the interception of the current data.

[0059] Then, the current data of the power frequency cycle before and after the starting point of the disturbance change is intercepted, and then the current data of the power frequency cycle before and after is used to detect arc faults, thereby achieving accurate current disturbance judgment. This method can realize fault detection for small current arcs, thereby filling the algorithm defects of the current traditional circuit breaker based on rigid threshold detection.

[0060] In an optional embodiment, using an improved CUSUM algorithm to perform current disturbance judgment on the current data includes the following steps:

[0061] Setting connected double-slide rail windows W1 and W2, and preprocessing the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window;

[0062] Obtain the distribution expectation threshold of the front window W1 and determine the offset threshold;

[0063] Based on the expected distribution value and the offset threshold, sequentially calculating the offset accumulation value of each data point of the preprocessed current data in the rear window W2, and iterating the offset accumulation value;

[0064] An offset limit is set, and the offset accumulation value after iteration is compared with the offset limit. If the comparison condition is met, it is determined that a current disturbance occurs in the rear window W2.

[0065] like Figure 1 As shown, Figure 1The vertical axis is the normalized current value, and the horizontal axis is the discrete current data (n). The connected double sliding track windows W1 and W2 are set on the graph, where the front window W1 is used to obtain the distribution expected value E and determine the positive offset threshold UB and the negative offset threshold LB; the rear window W2 is used to calculate the offset accumulation value of the data sequence in the window.

[0066] In this embodiment, the determination of the current disturbance is completed by comparing the current data in the front window W1 with the current data in the rear window W2.

[0067] The lengths of the two windows are both set to M, where the window length corresponds to M power frequency cycles.

[0068] In the initial case, the offset accumulation value of the initial post-window W2 is not set to the initial positive offset accumulation value. The accumulated negative offset value is Based on the distribution expected value E and the determined upper offset threshold UB and lower offset threshold LB, the offset accumulation value of each data point of the preprocessed current data in the rear window W2 is calculated in sequence, and the offset accumulation value is iterated.

[0069] Furthermore, the above process is shown in formula (1) to formula (3):

[0070]

[0071] Among them, if the kth data point x k When the upper or lower offset threshold is exceeded, the corresponding positive offset accumulation value Increase or negative offset accumulation value Set the deviation limit DL (Deviation Limit). When the accumulated deviation value C satisfies any one of the terms in formula (3), it can be considered that a disturbance change has been detected in the current in W2.

[0072] In an optional embodiment, preprocessing the current data intercepted by the double-slide window includes the following steps:

[0073] Calculate the effective value of the current data captured by the front window W1 and the rear window W2;

[0074] The average value of the effective value of the current data captured by the front window W1 is used as the reference value, and the effective values ​​of the current data captured by the front window W1 and the rear window W2 are calculated respectively to obtain the current normalization value.

[0075] It should be noted that in actual situations, the load combination is complex and the amplitude of the line current has no defined boundary. It is very inefficient to directly set a disturbance threshold. Therefore, in this embodiment, the current data intercepted by the double-slide window needs to be preprocessed.

[0076] Specifically, first, the effective values ​​of the current data intercepted by the front window W1 and the rear window W2 are calculated respectively, and the average value of the effective value of the current data intercepted by the front window W1 is used as the reference value. The effective values ​​of the current data intercepted by the front window W1 and the rear window W2 are calculated respectively to obtain the current normalized value of the front window and the current normalized value of the rear window respectively.

[0077] Taking the front window as an example, the calculation process of the current normalization value is shown in formula (4):

[0078]

[0079] Among them, I pu is the normalized value of current, I rms A valid value for the front window.

[0080] In an optional embodiment, after determining that a current disturbance occurs in the rear window W2, the method further includes: the double-track window slides and shifts Y positions, wherein M<Y<2M, and M is the window length of the front window W1 and the rear window W2;

[0081] After the double-slide window slides and shifts Y positions, the offset accumulation value of the rear window W2 is initialized.

[0082] It should be noted that after completing the data detection in the double-track window, the double-track window slides and shifts Y positions to detect the next data sequence. Among them, M<Y<2M means that the length of the sliding displacement is greater than the length of one window and less than the length of two windows. The purpose of this setting is to avoid repeated detection or missed detection.

[0083] Furthermore, in this embodiment, the current data X, the distribution expected value E, the upper offset threshold UB, the lower offset threshold LB and the offset limit DL are specifically set as shown in formulas (5) to (8):

[0084] (5)X=I pu

[0085] (6) E = 1

[0086] (7) UB = LB = 2%

[0087] (8)DL=λσ1

[0088] Among them, λ is a constant, and σ1 is the square root of the variance of the per-unit value sequence obtained by the front window W1.

[0089] In an optional embodiment, based on the current disturbance judgment result, locating the current disturbance starting point by the standard deviation and kurtosis of the current data comprises the following steps:

[0090] All current normalized values ​​in the front window W1 and the rear window W2 are constructed into a current normalized value sequence;

[0091] Overlapping and sorting the current normalized value sequence to construct an effective value matrix, and calculating the standard deviation and kurtosis in each part of the current normalized value sequence through the effective value matrix to obtain a standard deviation sequence and a kurtosis sequence;

[0092] The data point with the largest standard deviation is selected from the standard deviation sequence as the center point, and k kurtosis values ​​are extracted from the kurtosis sequence with the center point as the center. If the kurtosis value corresponding to the center point is the local minimum of the k kurtosis values, then the center point is the starting point of the current disturbance.

[0093] It should be noted that although the improved CUSUM algorithm can realize the recognition of current disturbance, since the algorithm needs to continuously accumulate the obtained offset values, there is an obvious time lag when obtaining the judgment result of the current disturbance, and it is necessary to further locate the starting point of the current disturbance. In this embodiment, the current normalized value sequences in the W1 and W2 windows are overlapped and sorted to construct the effective value matrix Mat pu , effective value matrix Mat pu As shown in formula (9):

[0094]

[0095] Among them, M+1 represents the total amount of normalized value data in the sequence, and k represents that k adjacent current normalized values ​​are extracted each time to form Mat pu of a line.

[0096] By Mat pu Calculate the standard deviation and kurtosis value in each part of the sequence. The standard deviation and kurtosis values ​​are shown in formula (10) and formula (11):

[0097]

[0098] Among them, Std n represents the standard deviation of the normalized current value of the nth row, Kur n represents the kurtosis value calculated from the normalized current value of the nth row, and both are used to reflect the degree of local disturbance in the normalized value sequence.

[0099] The standard deviation and kurtosis values ​​in formula (10) and formula (11) are combined to obtain the standard deviation sequence DS (Disturbance Sequence) and the kurtosis sequence KS (Kurtosis sequence). The standard deviation sequence DS and the kurtosis sequence KS are shown in formula (12) and formula (13):

[0100] (12)DS=[Std1 Std2 … Std M-k+2 ] T

[0101] (13)KS=[Kur1 Kur2 … Kur M-k+2 ] T

[0102] like Figure 2 As shown in the figure, the maximum value of the standard deviation will appear at the starting point of the current disturbance. The point corresponding to the maximum value of the standard deviation will be taken as the center point, and a total of k kurtosis values ​​before and after the center point will be extracted. If the kurtosis value corresponding to the center point is the local minimum among the k kurtosis values, then the center point is recorded as the starting point of the current disturbance change.

[0103] In an optional embodiment, arc fault detection is performed on current data of previous and next power frequency cycles using even harmonic technology, including the following steps:

[0104] The current data of the power frequency cycles before and after the front window W1 and the rear window W2 are processed by short-time Fourier transform to obtain the front even harmonic sum sequence and the rear even harmonic sum sequence;

[0105] Calculating the even harmonic sum ratio of the sum of the preceding even harmonic sum sequence and the sum of the following even harmonic sum sequence;

[0106] A detection threshold is set, and if the even harmonic sum ratio is greater than the detection threshold, an arc fault exists.

[0107] It should be noted that the introduction of even harmonics and A even This characteristic quantity, where n represents the order, m represents the upper limit of the order, and A 2n In this embodiment, m is set to 50, and the sum of even harmonics within 5 kHz is calculated using formula (14):

[0108]

[0109] Considering the fact that arc faults will last for a long time after the fault occurs, the current data of 20 power frequency cycles before and after the starting point of the current disturbance are intercepted, and the corresponding even harmonics and sequences are obtained by short-time Fourier transform STFT. The leading even harmonics and sequences are shown in formula (15), and the trailing even harmonics and sequences are shown in formula (16):

[0110]

[0111] Among them, norm represents the data obtained in W1, arc represents the data obtained in W2, and the subscripts 1 to 20 represent the 1st to 20th power frequency cycles respectively.

[0112] Then, the sum of the even harmonics and the sequence is calculated using formula (17), and the even harmonic sum ratio of the sum of the leading even harmonics and the sequence and the sum of the trailing even harmonics and the sequence is calculated using formula (18):

[0113]

[0114] The technicians in this field can set S according to the actual situation. set As the detection threshold, when the even harmonics and the ratio S Ratio Greater than S set , it is considered that an arc fault has occurred.

[0115] The small current arc fault detection method based on time-frequency domain disturbance provided in this embodiment can be applied to Figure 3 (a) and Figure 3 In the two fault detection methods shown in (b), only data with a relatively low acquisition frequency (below 10kHz) and a short time scale (within 50ms) are needed to detect arc faults in low-voltage AC distribution systems.

[0116] Example 2

[0117] Embodiment 2 of the present invention proposes a small current arc fault detection system based on time-frequency domain disturbance, including:

[0118] A current disturbance module is used to obtain current data and use an improved CUSUM algorithm to perform current disturbance judgment on the current data;

[0119] A starting point module, used for locating the starting point of the current disturbance through the standard deviation and kurtosis of the current data based on the current disturbance judgment result;

[0120] The fault detection module is used to intercept the current data of the power frequency cycle before and after the starting point of the current disturbance, and use the even harmonic technology to perform arc fault detection on the current data of the power frequency cycle before and after to obtain the detection result.

[0121] In an optional embodiment, the current disturbance module includes:

[0122] A preprocessing unit, used to set the connected double-slide rail windows W1 and W2, and preprocess the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window;

[0123] A front window unit, used to obtain a distribution expectation threshold of the front window W1 and determine an offset threshold;

[0124] A rear window unit, used to sequentially calculate the offset accumulation value of each data point of the preprocessed current data in the rear window W2 based on the distribution expected value and the offset threshold, and iterate the offset accumulation value;

[0125] The comparison unit is used to set an offset limit, compare the offset accumulation value after iteration with the offset limit, and if the comparison condition is met, determine that a current disturbance occurs in the rear window W2.

[0126] Example 3

[0127] Figure 4 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention is shown in FIG. Figure 4 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the computer device can be one or more. Figure 4 A processor 21 is taken as an example; the processor 21, the memory 22, the input device 23 and the output device 24 in the electronic device can be connected by a bus or other means. Figure 4 The example of connecting through bus is taken in the following.

[0128] The memory 22 is a computer-readable storage medium that can be used to store software programs, computer executable programs, and modules. The processor 21 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 22, that is, the small current arc fault detection method based on time-frequency domain disturbance of Example 1 is implemented.

[0129] The memory 22 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 22 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 22 may further include a memory remotely arranged relative to the processor 21, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0130] The input device 23 can be used to receive the ID and password input by the user, etc. The output device 24 is used to output the network configuration page.

[0131] Example 4

[0132] Embodiment 4 of the present invention further provides a computer-readable storage medium, wherein the computer-executable instructions are used to implement the small current arc fault detection method based on time-frequency domain disturbance as provided in Embodiment 1 when executed by a computer processor.

[0133] An embodiment of the present invention provides a storage medium containing computer executable instructions, and its computer executable instructions are not limited to the method operations provided in Example 1, and can also execute related operations in the small current arc fault detection method based on time-frequency domain disturbance provided in any embodiment of the present invention.

[0134] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A small current arc fault detection method based on time-frequency domain disturbance, characterized in that: The steps include: Acquire current data, and use an improved CUSUM algorithm to perform current disturbance judgment on the current data; Based on the current disturbance judgment result, locating the current disturbance starting point through the standard deviation and kurtosis of the current data; The current data of the power frequency cycles before and after the starting point of the current disturbance are intercepted, and the arc fault detection is performed on the current data of the power frequency cycles before and after using the even harmonic technology to obtain the detection result.

2. The small current arc fault detection method based on time-frequency domain disturbance according to claim 1 is characterized in that: The improved CUSUM algorithm is used to perform current disturbance judgment on the current data, including the following steps: Setting connected double-slide rail windows W1 and W2, and preprocessing the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window; Obtain the distribution expectation threshold of the front window W1 and determine the offset threshold; Based on the expected distribution value and the offset threshold, sequentially calculating the offset accumulation value of each data point of the preprocessed current data in the rear window W2, and iterating the offset accumulation value; An offset limit is set, and the offset accumulation value after iteration is compared with the offset limit. If the comparison condition is met, it is determined that a current disturbance occurs in the rear window W2.

3. The small current arc fault detection method based on time-frequency domain disturbance according to claim 2 is characterized in that: Preprocessing the current data intercepted by the double-slide window includes the following steps: Calculate the effective value of the current data captured by the front window W1 and the rear window W2; The average value of the effective value of the current data captured by the front window W1 is used as the reference value, and the effective values ​​of the current data captured by the front window W1 and the rear window W2 are calculated respectively to obtain the current normalization value.

4. The small current arc fault detection method based on time-frequency domain disturbance according to claim 2 is characterized in that: After determining that a current disturbance occurs in the rear window W2, the method further includes: sliding the double-track window to Y positions, wherein M<Y<2M, and M is the window length of the front window W1 and the rear window W2; After the double-slide window slides and shifts Y positions, the offset accumulation value of the rear window W2 is initialized.

5. The small current arc fault detection method based on time-frequency domain disturbance according to claim 3 is characterized in that: Based on the current disturbance judgment result, the current disturbance starting point is located by the standard deviation and kurtosis of the current data, including the following steps: All current normalized values ​​in the front window W1 and the rear window W2 are constructed into a current normalized value sequence; Overlapping and sorting the current normalized value sequence to construct an effective value matrix, and calculating the standard deviation and kurtosis in each part of the current normalized value sequence through the effective value matrix to obtain a standard deviation sequence and a kurtosis sequence; The data point with the largest standard deviation is selected from the standard deviation sequence as the center point, and k kurtosis values ​​are extracted from the kurtosis sequence with the center point as the center. If the kurtosis value corresponding to the center point is the local minimum of the k kurtosis values, then the center point is the starting point of the current disturbance.

6. The small current arc fault detection method based on time-frequency domain disturbance according to claim 2 is characterized in that: The arc fault detection is performed on the current data of the previous and next power frequency cycles using even harmonic technology, including the following steps: The current data of the power frequency cycles before and after the front window W1 and the rear window W2 are processed by short-time Fourier transform to obtain the front even harmonic sum sequence and the rear even harmonic sum sequence; Calculating an even harmonic sum ratio of the sum of the preceding even harmonic sum sequence and the sum of the following even harmonic sum sequence; A detection threshold is set, and if the even harmonic sum ratio is greater than the detection threshold, an arc fault exists.

7. A small current arc fault detection system based on time-frequency domain disturbance, characterized in that: include: A current disturbance module is used to obtain current data and use an improved CUSUM algorithm to perform current disturbance judgment on the current data; A starting point module, used for locating the starting point of the current disturbance through the standard deviation and kurtosis of the current data based on the current disturbance judgment result; The fault detection module is used to intercept the current data of the power frequency cycle before and after the starting point of the current disturbance, and use the even harmonic technology to perform arc fault detection on the current data of the power frequency cycle before and after to obtain the detection result.

8. The low current arc fault detection system based on time-frequency domain disturbance according to claim 7 is characterized in that: The current disturbance module comprises: A preprocessing unit, used to set the connected double-slide rail windows W1 and W2, and preprocess the current data intercepted by the double-slide rail windows to obtain preprocessed current data; wherein W1 is the front window and W2 is the rear window; A front window unit, used to obtain a distribution expectation threshold of the front window W1 and determine an offset threshold; A rear window unit, used to sequentially calculate the offset accumulation value of each data point of the preprocessed current data in the rear window W2 based on the distribution expected value and the offset threshold, and iterate the offset accumulation value; The comparison unit is used to set an offset limit, compare the offset accumulation value after iteration with the offset limit, and if the comparison condition is met, determine that a current disturbance occurs in the rear window W2.

9. An electronic device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for detecting a small current arc fault based on time-frequency domain disturbance as claimed in any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for detecting a small current arc fault based on time-frequency domain disturbance as claimed in any one of claims 1 to 6 is implemented.

Images