Arcing fault detection method, apparatus, and system

Abstract

This invention discloses a method, device, and system for arcing fault detection. The method includes: dividing and aggregating the actual electrical signals of the entire range of the electrical equipment under test according to the characteristics of the electrical signals, obtaining aggregated sub-interval actual electrical signals; determining the optimal aggregated sub-interval based on the energy spectral density of the actual electrical signals of each aggregated sub-interval in a preset multi-frame electrical signal, and an energy spectral density threshold; inputting the actual electrical signals of the aggregated sub-interval corresponding to the optimal aggregated sub-interval into a target arcing detection model that matches the optimal aggregated sub-interval, obtaining an actual arcing probability value; and determining whether the electrical equipment under test has an arcing fault based on the comparison result between the actual arcing probability value and the preset arcing probability threshold. This invention solves the technical problem of low accuracy in arcing fault detection in the prior art, effectively improving the accuracy of arcing probability value determination, and balancing detection accuracy with equipment operational stability.

Description

Technical Field

[0001] This invention relates to the field of electrical safety testing technology, and in particular to a method, equipment and system for detecting arcing faults. Background Technology

[0002] Arcing faults are a common and potentially dangerous phenomenon during the operation of electrical equipment. Arcing is accompanied by high temperatures, intense light, and electromagnetic interference, potentially leading to serious consequences such as fires and equipment damage. Therefore, accurate arc detection is crucial for ensuring the safety of electrical systems. Traditional arc detection techniques are susceptible to interference from inverter high-frequency noise, load variations, and external environmental factors. This interference can cause the characteristics of electrical signal fluctuations during normal operation to overlap with those generated by arcing, frequently leading to misjudgments. Normal fluctuations may be mistaken for arcing, or, when a true arc occurs, the interference may mask the signal characteristics, resulting in missed detection. Furthermore, the frequency of arcing signals is not fixed but varies depending on the type of arcing, its cause, and the operating conditions of the photovoltaic system. Fixed-bandwidth filters cannot flexibly adapt to these frequency changes and may miss arcing signals within certain frequency ranges, resulting in incomplete detection results and difficulty in accurately and promptly identifying potential arcing faults. Summary of the Invention

[0003] This invention provides a method, device, and system for detecting arcing faults, in order to solve the technical problem of low accuracy in arcing fault detection in the prior art.

[0004] According to one aspect of the present invention, an arcing fault detection method is provided, the method comprising:

[0005] Based on the characteristics of electrical signals, the actual electrical signals of the entire range of the electrical equipment under test are divided and aggregated to obtain the aggregated sub-range actual electrical signals.

[0006] The optimal aggregation sub-interval is determined based on the energy spectral density of the actual electrical signal in each aggregation sub-interval in a preset multi-frame electrical signal, and the energy spectral density threshold.

[0007] The actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval is input into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value;

[0008] The presence of an arcing fault in the electrical equipment under test is determined based on the comparison between the actual arcing probability value and the preset arcing probability threshold.

[0009] According to another aspect of the present invention, an arcing fault detection device is provided, the arcing fault detection device comprising: a signal acquisition unit, a signal processor, and a microprocessor; the output terminal of the signal acquisition unit is connected to the input terminal of the signal processor, and the output terminal of the signal processor is connected to the microprocessor;

[0010] The signal acquisition unit is used to acquire the original full-range actual electrical signal of the electrical equipment under test, and sends the original full-range actual electrical signal to the signal processor. The signal processor performs filtering and amplification operations on the original full-range actual electrical signal to obtain the actual full-range actual electrical signal. The microprocessor divides and aggregates the full-range actual electrical signal of the electrical equipment under test according to the characteristics of the electrical signal to obtain aggregated sub-interval actual electrical signals. The optimal aggregated sub-interval is determined according to the energy spectral density of each aggregated sub-interval actual electrical signal in a preset multi-frame electrical signal and the energy spectral density threshold. The actual electrical signal of the aggregated sub-interval corresponding to the optimal aggregated sub-interval is input into the target arc detection model that matches the optimal aggregated sub-interval to obtain the actual arc probability value. The actual arc probability value and the preset arc probability threshold are compared to determine whether the electrical equipment under test has an arc fault.

[0011] According to another aspect of the present invention, an arcing fault detection system is provided, the arcing fault detection system comprising: an electrical device under test and an arcing fault detection device as described in any embodiment of the present invention.

[0012] The technical solution of this invention dynamically divides the actual electrical signal of the entire range of the electrical equipment under test according to the characteristics of the electrical signal, ensuring that the divided range can completely capture local telephone features and avoid redundant information interference. Then, the divided ranges are aggregated, which makes the arcing features of the aggregated sub-ranges more concentrated and reduces noise interference. After dynamic division and aggregation, the fluctuation of false alarm / false alarm scores can be significantly reduced, and the stability of feature extraction can be improved. Furthermore, the optimal aggregated sub-range is determined according to the energy spectral density of the actual electrical signal of each aggregated sub-range in a preset multi-frame electrical signal and the energy spectral density threshold, which can effectively improve the selection probability of the optimal range. Moreover, the actual electrical signal of the aggregated sub-range corresponding to the optimal aggregated sub-range is input into the target arcing detection model that matches the optimal aggregated sub-range to obtain the actual arcing probability value. This can accurately quantify the matching degree between the signal and the arcing features, avoid the problem of insufficient adaptability of general models, and effectively improve the accuracy of arcing probability value determination. Furthermore, by comparing the actual arcing probability value with the preset arcing probability threshold, the detection accuracy and equipment operation stability can be effectively balanced.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart of an arc fault detection method provided in an embodiment of the present invention;

[0016] Figure 2 This is a flowchart of another arc fault detection method provided in an embodiment of the present invention;

[0017] Figure 3 This is a flowchart of another arc fault detection method provided in an embodiment of the present invention;

[0018] Figure 4 This is a flowchart of another arc fault detection method provided in an embodiment of the present invention;

[0019] Figure 5 This is a block diagram of an arc fault detection system provided in an embodiment of the present invention;

[0020] Figure 6 This is a schematic diagram of a frequency domain waveform with a relatively light load under a relatively large current, provided by an embodiment of the present invention;

[0021] Figure 7 This is a schematic diagram of a frequency domain waveform with a relatively small current and a relatively light load, provided by an embodiment of the present invention;

[0022] Figure 8 This is a schematic diagram of a frequency domain waveform with a relatively small current and a relatively heavy load, provided by an embodiment of the present invention.

[0023] Figure 9 This is a flowchart provided by an embodiment of the present invention, showing the process from acquiring raw electrical signals across the entire range to detecting arcing faults.

[0024] Figure 10 This is a schematic diagram illustrating the implementation of initial sub-interval division and arc probability value according to an embodiment of the present invention;

[0025] Figure 11 This is a schematic diagram illustrating the implementation of aggregator sub-interval partitioning and optimal aggregator sub-interval determination provided by an embodiment of the present invention;

[0026] Figure 12 This is a schematic diagram illustrating the display of the energy spectral density of multiple frames of electrical signals according to an embodiment of the present invention;

[0027] Figure 13 This is a flowchart illustrating the implementation of model optimization and arc detection protection provided in an embodiment of the present invention;

[0028] Figure 14 This is a schematic diagram illustrating the false alarms and false negatives of arc detection using a traditional interval model, as provided in an embodiment of the present invention.

[0029] Figure 15 This is a schematic diagram illustrating the false alarms and false negatives of arc detection using an optimized target arc detection model, as provided in an embodiment of the present invention.

[0030] Figure 16 This is a schematic diagram of the structure of an arc fault detection device provided in an embodiment of the present invention;

[0031] Figure 17 This is a schematic diagram of another arc fault detection device provided in an embodiment of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] In one embodiment, Figure 1This is a flowchart of an arcing fault detection method provided by an embodiment of the present invention. This embodiment is applicable to detecting whether an arcing fault occurs during the operation of electrical equipment. The method can be executed by an arcing fault detection device, which can be implemented in hardware and / or software and can be configured in an arcing fault detection equipment. Figure 1 As shown, the method includes:

[0035] S110. Based on the characteristics of the electrical signal, divide and aggregate the actual electrical signal of the entire range of the electrical equipment under test to obtain the aggregated sub-range actual electrical signal.

[0036] In one example, the electrical equipment under test (EDT) refers to electrical equipment that needs to be tested for arcing faults. This is used to investigate the arcing risk of the voltage / current signals generated during the operation of the EDT, ensuring its safe and stable operation. For example, the EDT may include, but is not limited to, at least one of the following scenarios: industrial electrical equipment (e.g., motors, frequency converters, transmission lines, etc. in a factory production line) and civil electrical equipment (e.g., sockets, switches, etc.). The electrical signal refers to the voltage or current signal collected during the operation of the EDT. The electrical signal characteristics refer to the changing patterns of the voltage or current signals collected during the operation of the EDT, used for dynamic differentiation, feature extraction, and arcing detection. For example, electrical signal characteristics may include at least one of the following dimensions: signal stability, signal frequency distribution characteristics, and signal energy concentration characteristics. For instance, from the perspective of signal stability, this may include: relatively stable signal changes, large signal fluctuations, and signals under complex operating conditions.

[0037] In one example, the full-range actual electrical signal refers to the complete voltage or current signal collected from the electrical equipment under test, which has undergone signal preprocessing and has not been segmented into ranges, and can cover the entire time domain range within the detection period; the aggregated sub-range actual electrical signal refers to the actual electrical signal obtained by splitting the full-range actual electrical signal into multiple initial sub-range actual electrical signals and then merging the multiple initial sub-range actual electrical signals through a clustering algorithm. It is an aggregation and integration of initial sub-range actual electrical signals with similar arcing characteristics.

[0038] In this embodiment, the stability of the signal can be used to distinguish and divide the actual electrical signal of the entire range of the electrical equipment under test, resulting in multiple initial sub-range actual electrical signals; then, the multiple initial sub-range actual electrical signals are merged according to the clustering algorithm to obtain multiple aggregated sub-range actual electrical signals.

[0039] It should be noted that the number of aggregated sub-intervals is less than or equal to the number of initial sub-intervals. For example, suppose the actual electrical signal of the entire interval is divided into N initial sub-intervals; then a clustering algorithm is used to merge the N initial sub-intervals to obtain M aggregated sub-intervals, where M is a positive integer less than or equal to N, and N is a positive integer greater than or equal to 2.

[0040] S120. Determine the optimal aggregation sub-interval based on the energy spectral density of the actual electrical signal in each aggregation sub-interval within a preset multi-frame electrical signal, and the energy spectral density threshold.

[0041] In one example, energy spectral density is used to quantify the distribution of energy of the actual electrical signal in the aggregated sub-interval at different frequencies, to reflect the degree of energy concentration of the actual electrical signal in the aggregated sub-interval within the interval corresponding to the preset multi-frame electrical signal, and to determine whether the aggregated sub-interval contains arcing characteristics; the optimal aggregated sub-interval refers to the set of intervals with the most significant arcing characteristics and the most concentrated energy selected from the multiple aggregated sub-intervals; the preset multi-frame electrical signal is a set of pre-processed electrical signals of continuous and fixed duration collected to verify the energy stability of the aggregated sub-interval, and is used to reduce misjudgments caused by noise interference from single-frame signals.

[0042] In this embodiment, a microprocessor can acquire multiple frames of preprocessed electrical signals in succession, with the duration of each frame of electrical signal being consistent with the duration of the full-range signal. Then, a Fourier transform can be performed on the actual electrical signal within each aggregated sub-range to obtain the frequency domain signal corresponding to that aggregated sub-range. Based on the frequency domain signal and the duration of the corresponding aggregated sub-range, the energy spectral density of the corresponding aggregated sub-range in each frame of the multi-frame continuous preprocessed electrical signal can be determined. Then, multiple aggregated sub-ranges with an average energy spectral density greater than or equal to an energy spectral density threshold are constituted as the optimal sub-range set. Finally, high-energy feature regions are selected from the optimal sub-range set as the optimal aggregated sub-ranges.

[0043] S130. Input the actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value.

[0044] In one embodiment, the process of constructing the target arc detection model includes: dividing and aggregating the full-range training electrical signals in the electrical signal training set according to the characteristics of the electrical signals to obtain aggregated sub-range training electrical signals; determining the optimal aggregated sub-range based on the energy spectral density of each aggregated sub-range training electrical signal in a preset multi-frame electrical signal and the energy spectral density threshold; and training the initial arc detection model based on the aggregated sub-range training electrical signals corresponding to the optimal aggregated sub-range to obtain the target arc detection model.

[0045] In one example, the full-range training electrical signal refers to the complete voltage or current signal that has been pre-acquired for training and has undergone signal preprocessing, as well as the signal that has not been segmented into ranges, which can cover the entire time domain range within the detection period.

[0046] It should be noted that the process of dividing and aggregating the training electrical signals across the entire range, as well as the process of optimally aggregating sub-ranges, can be found in the explanation of the above embodiments, and will not be repeated here.

[0047] In this embodiment, the target arcing detection model refers to a pre-created and trained arcing detection model whose detection accuracy meets user requirements; the target arcing detection model can be a convolutional neural network model. It should be noted that, since the frequency domain waveforms of electrical signals have different distinguishability across different frequency bands under different currents and / or different loads, in order to facilitate accurate determination of whether the electrical equipment under test has experienced an arcing fault based on the actual electrical signals within the aggregate sub-interval, different target arcing fault detection models can be trained and constructed for different currents and / or different loads.

[0048] In one example, the actual arcing probability value reflects the real-time likelihood of an arcing fault occurring in the electrical equipment under test under its current operating state. It is the probability value output by the final optimized target arcing detection model. Generally, the actual arcing probability value is a quantified result of the matching between signal features and arcing fault features, and can be a value between 0 and 1. The larger the value, the greater the probability of arcing.

[0049] In this embodiment, a suitable target arc detection model can be selected based on the characteristics of the frequency domain waveform of the actual electrical signal in the optimal aggregate sub-interval, and the actual electrical signal in the optimal aggregate sub-interval can be input into the target arc detection model to obtain the corresponding actual arc probability value.

[0050] S140. Determine whether the electrical equipment under test has an arcing fault based on the comparison between the actual arcing probability value and the preset arcing probability threshold.

[0051] In one example, the arcing fault detection result is used to characterize whether the electrical equipment under test has an arcing fault in its current operating state; the preset arcing probability threshold is a preset value used to quantify and distinguish between arcing faults, suspected arcing, and normal states, and can determine the arcing detection sensitivity, false alarm rate, and false negative rate. In an embodiment, the preset arcing probability threshold can be combined to classify the arcing fault detection result into arcing faults, suspected arcing, and normal states.

[0052] In one embodiment, determining whether the electrical equipment under test has an arcing fault based on the comparison between the actual arcing probability value and a preset arcing probability threshold includes: if the actual arcing probability value is less than a first actual arcing probability threshold, determining that the electrical equipment under test has an arcing fault; if the actual arcing probability value is greater than or equal to the first actual arcing probability threshold and less than a second actual arcing probability threshold, determining that the electrical equipment under test is suspected of having an arcing fault; and if the actual arcing probability value is greater than or equal to the second actual arcing probability threshold, determining that the electrical equipment under test has not had an arcing fault.

[0053] In one example, the first actual arcing probability threshold is less than the second actual arcing probability threshold. The first actual arcing probability threshold is a threshold value used to characterize whether an arcing fault has occurred, which can also be understood as a threshold value for whether a level-one alarm has occurred. The second actual arcing probability threshold is a threshold value used to characterize whether a suspected arcing fault has occurred, which can also be understood as a threshold value for whether a level-two alarm has occurred. When the actual arcing probability value is less than the first actual arcing probability threshold, the actual electrical signal corresponding to the electrical equipment under test has clear arcing characteristics with a very high probability, and an arcing fault has occurred, which may lead to damage to the electrical equipment under test or a safety accident. When the actual arcing probability value is greater than or equal to the first actual arcing probability threshold and less than the second actual arcing probability threshold, the actual electrical signal corresponding to the electrical equipment under test has some arcing characteristics, but does not meet the criteria for a clear fault; it may be noise interference or initial minor arcing. When the actual arcing probability value is greater than or equal to the second actual arcing probability threshold, the actual electrical signal corresponding to the electrical equipment under test has no arcing characteristics, meaning the operating state of the electrical equipment under test is stable, and the signal characteristics are consistent with the non-arcing frequency domain waveform.

[0054] The technical solution of this embodiment dynamically divides the actual electrical signal of the entire range of the electrical equipment under test according to the characteristics of the electrical signal, ensuring that the divided range can completely capture local telephone features and avoid redundant information interference. Then, the divided ranges are aggregated, which makes the arcing features of the aggregated sub-ranges more concentrated and reduces noise interference. After dynamic division and aggregation, the fluctuation of false alarm / false alarm scores can be significantly reduced, and the stability of feature extraction can be improved. Furthermore, the optimal aggregated sub-range is determined according to the energy spectral density of the actual electrical signal of each aggregated sub-range in a preset multi-frame electrical signal and the energy spectral density threshold, which can effectively improve the selection probability of the optimal range. In addition, the actual electrical signal of the aggregated sub-range corresponding to the optimal aggregated sub-range is input into the target arcing detection model that matches the optimal aggregated sub-range to obtain the actual arcing probability value. This can accurately quantify the matching degree between the signal and the arcing features, avoid the problem of insufficient adaptability of general models, and effectively improve the accuracy of probability value determination. Furthermore, by comparing the actual arcing probability value with the preset arcing probability threshold, the detection accuracy and equipment operation stability can be effectively balanced.

[0055] In one embodiment, Figure 2 This is a flowchart of another arc fault detection method provided by an embodiment of the present invention. This embodiment further refines the process of determining the actual electrical signals of the aggregate sub-interval based on the above embodiments. Figure 2 As shown, the method includes:

[0056] S210. Based on the characteristics of the electrical signal, divide the actual electrical signal of the entire range of the electrical equipment under test into multiple initial sub-range actual electrical signals.

[0057] In one embodiment, S210 includes: if the electrical signal changes smoothly, dividing the actual electrical signal of the entire range of the electrical equipment under test into intervals according to fixed time intervals to obtain multiple initial sub-interval actual electrical signals; if the electrical signal has sudden peaks, dynamically dividing the actual electrical signal of the entire range of the electrical equipment under test into intervals according to the points of change in electrical signal intensity to obtain multiple initial sub-interval actual electrical signals; if the electrical signal does not have a fixed change pattern, dynamically dividing the actual electrical signal of the entire range of the electrical equipment under test into intervals using an adaptive time window division method to obtain multiple initial sub-interval actual electrical signals.

[0058] In one example, a stable electrical signal means that the signal amplitude fluctuates little, there are no sudden peaks, and the frequency distribution is uniform; a sudden peak in an electrical signal means that the signal fluctuates greatly, there are obvious abrupt changes in signal strength, and the peak value fluctuates frequently; an electrical signal does not have a fixed pattern of change, meaning that under complex operating conditions, the signal is greatly affected by the magnitude of the current and the type of the load, and the signal entropy fluctuates drastically.

[0059] In one example, if the actual electrical signal of the electrical equipment under test is stable throughout the entire range, the actual electrical signal of the electrical equipment under test can be divided into multiple initial sub-range actual electrical signals at fixed time intervals. Specifically, it can be divided according to a preset fixed time length, for example, each initial sub-range is 0.002 seconds, and the length of each initial sub-range is uniform, so that the interval division process can be completed efficiently.

[0060] In one example, if there is a sudden peak in the electrical signal, the actual electrical signal of the entire range of the electrical equipment under test is dynamically divided into multiple initial sub-range actual electrical signals based on the change point of the electrical signal strength. Specifically, the peak value of the derivative of the signal strength (i.e. the point of sudden change) can be used as the dividing boundary to accurately split the feature change region and avoid the arcing feature being truncated across the range.

[0061] In one example, if the electrical signal does not have a fixed pattern of change, an adaptive time window division method is used to dynamically divide the actual electrical signal of the entire range of the electrical equipment under test into multiple initial sub-range actual electrical signals. Specifically, the window length can be shortened when the signal entropy is high (i.e., the features are complex) and extended when the signal entropy is low (i.e., the features are simple), so as to adapt to the changes in the signal.

[0062] In this embodiment, by decomposing the actual electrical signal of the entire range into multiple local initial sub-range actual electrical signals, the arcing characteristics can be separated from the complex information of the entire range, avoiding feature overload caused by the excessive signal range. Furthermore, by adaptively dividing the entire range, it can adapt to model changes under different operating conditions, effectively avoiding the loss of features such as low-frequency arcing characteristics under low current heavy loads or redundant interference such as noise ranges without arcing characteristics caused by fixed division, thereby improving the detection accuracy of whether arcing faults exist.

[0063] S220. Input the actual electrical signal of each initial sub-interval into the initial arc detection model to obtain the arc probability value of the actual electrical signal of the corresponding initial sub-interval.

[0064] In one example, the initial arc detection model is organized by using preprocessed full-range electrical signals as training data input into the model, and training the initial detection model by learning the frequency domain characteristics and energy distribution characteristics of arc signals and non-arc signals.

[0065] In this embodiment, the actual electrical signal of each initial sub-interval can be input into the initial arc detection model. The initial arc detection model can calculate the matching degree between the extracted model features of the initial sub-interval and the arc feature template learned during the training process, and convert the matching degree into an arc probability value through quantization. The initial arc detection model outputs the arc probability value corresponding to each initial sub-interval, forming a complete mapping relationship between the initial sub-interval and the arc probability value.

[0066] S230. Based on the arcing probability value of the actual electrical signal in the initial sub-interval, and using a clustering algorithm to aggregate the actual electrical signals of multiple initial sub-intervals, multiple aggregated sub-interval actual electrical signals are obtained.

[0067] In one embodiment, S230 includes: determining an average arcing probability based on the arcing probability value of the actual electrical signal in each initial sub-interval and the number of initial sub-intervals;

[0068] The first aggregation arcing probability threshold and the second aggregation arcing probability threshold are determined according to the average arcing probability; wherein, the first aggregation arcing probability threshold is less than the second aggregation arcing probability threshold;

[0069] The actual electrical signals of multiple initial sub-intervals are aggregated based on the first and second aggregated arc probability thresholds to obtain multiple aggregated sub-interval actual electrical signals.

[0070] In one example, the first aggregation arc probability threshold and the second aggregation arc probability threshold are used as threshold values ​​for merging multiple initial sub-intervals.

[0071] In one example, the arcing probability values ​​of the actual electrical signals in each initial sub-interval can be summed to obtain a total arcing probability value. The ratio of this total arcing probability value to the number of initial sub-intervals is then used to obtain the average arcing probability. A first aggregated arcing probability threshold and a second aggregated arcing probability threshold are then set based on this average arcing probability. Specifically, the arcing probability values ​​of the actual electrical signals in the initial sub-intervals can be compared with the set dual thresholds (i.e., the first aggregated arcing probability threshold and the second aggregated arcing probability threshold) to divide the initial sub-intervals into a low-probability interval set, a medium-probability interval set, and a high-probability interval set. Then, in a continuous sequence, the arcing probability values ​​corresponding to two or more adjacent initial sub-intervals that simultaneously fall within the same probability interval set can be merged into a single aggregated sub-interval, resulting in multiple aggregated sub-interval actual electrical signals. Specifically, all initial sub-intervals with a probability less than the first aggregation arcing probability threshold are divided into a low-probability interval set, all initial sub-intervals with a probability greater than or equal to the first aggregation arcing probability threshold and less than or equal to the second aggregation arcing probability threshold are divided into a medium-probability interval set, and all initial sub-intervals with a probability greater than the second aggregation arcing probability threshold are divided into a high-probability interval set; then, the initial sub-intervals in each probability interval set are merged into multiple aggregation sub-intervals according to the principle of continuous merging.

[0072] S240. Determine the optimal aggregation sub-interval based on the energy spectral density of the actual electrical signal in each aggregation sub-interval within a preset multi-frame electrical signal, and the energy spectral density threshold.

[0073] S250. Input the actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value.

[0074] S260. Determine whether the electrical equipment under test has an arcing fault based on the comparison between the actual arcing probability value and the preset arcing probability threshold.

[0075] In the technical solution of the embodiment, based on the above embodiment, by first dividing the actual electrical signal of the entire interval into intervals to obtain multiple initial sub-interval actual electrical signals, the invalid calculation of a massive number of initial sub-intervals can be effectively avoided, and the efficiency of optimal aggregation sub-interval screening can be improved; then, the multiple initial sub-interval actual electrical signals are aggregated according to the first aggregation arcing probability threshold and the second aggregation arcing probability threshold to obtain multiple aggregated sub-interval actual electrical signals, which can completely cover the high-energy characteristic interval of the arcing signal, avoid the fragmentation of the optimal interval due to the initial sub-intervals being too fine, and thus ensure the integrity of the optimal interval.

[0076] In one embodiment, Figure 3 This is a flowchart of another arc fault detection method provided by an embodiment of the present invention. This embodiment further refines the process of determining the optimal aggregation sub-interval based on the above embodiments. Figure 3 As shown, this embodiment includes:

[0077] S310. Based on the characteristics of the electrical signal, divide and aggregate the actual electrical signal of the entire range of the electrical equipment under test to obtain the aggregated sub-range actual electrical signal.

[0078] S320. Determine the energy spectral density of the actual electrical signal in each aggregate sub-interval within the pre-acquired multi-frame electrical signals.

[0079] In one embodiment, S320 includes: performing a Fourier transform on the actual electrical signal of each aggregate sub-interval to obtain a frequency domain signal; and using the ratio between the square of the frequency domain signal and the duration of the corresponding aggregate sub-interval as the energy spectral density of the corresponding aggregate sub-interval.

[0080] In one example, the duration of each aggregated sub-interval can be the same or different. The duration of each aggregated sub-interval can be represented using different units such as time slots or milliseconds.

[0081] In one example, a Fourier transform can be performed on the actual electrical signal of each aggregate sub-interval to obtain the frequency domain signal; then the square of the absolute value of the frequency domain signal is determined, and the ratio between the square value and the duration of the corresponding aggregate sub-interval is taken as the energy spectral density of the corresponding aggregate sub-interval.

[0082] S330. Select all aggregate subintervals whose average energy spectral density is greater than or equal to the energy spectral density threshold as the optimal subinterval set.

[0083] In one example, the average energy spectral density refers to the average energy spectral density of each aggregated sub-interval across a preset number of frames of electrical signals. This can be understood as follows: the larger the average energy spectral density, the more concentrated the energy of the electrical signal within that aggregated sub-interval, and the more pronounced the arcing characteristics. In one example, an energy spectral density threshold can be pre-configured, and all aggregated sub-intervals with an average energy spectral density greater than or equal to the threshold can be included in the optimal sub-interval set. This ensures that all aggregated sub-intervals within the optimal sub-interval set can completely cover the high-energy characteristic region of the arcing signal.

[0084] S340. Use cross-validation to determine the optimal interval coefficient corresponding to each aggregated subinterval in the optimal subinterval set.

[0085] In one example, the interval coefficients corresponding to each aggregated sub-interval in the optimal sub-interval set can be initialized, and a portion of the electrical signal training set can be selected as a validation set to evaluate the effect of different combinations of interval coefficients. Different interval coefficients are applied to each aggregated sub-interval in the optimal sub-interval set, and each combination is applied to the target arc detection model through cross-validation and evaluated on the validation set. By comparing the detection effects under different coefficient combinations, the optimal combination of interval coefficients is selected. The evaluation metrics for the detection effect can include accuracy, recall, F1 score, etc.

[0086] S350. Using the optimal interval coefficient, select the aggregated sub-interval with the best detection effect from the set of optimal sub-intervals as the optimal aggregated sub-interval.

[0087] In one example, after determining the optimal combination of interval coefficients, the detection performance of each evaluation index under this optimal combination of interval coefficients can be determined. If performing single-objective optimization, for example, directly selecting the coefficient with the highest accuracy, then the aggregated sub-interval with the highest accuracy is taken as the optimal aggregated sub-interval; if performing multi-objective optimization, a weighted scoring method can be used to calculate the comprehensive score and then sort them, and the aggregated sub-interval with the highest comprehensive score is taken as the optimal aggregated sub-interval. For example, suppose the optimal sub-interval set contains three aggregated sub-intervals (aggregate sub-interval 1, aggregate sub-interval 2, and aggregate sub-interval 3), and the evaluation metrics are precision, recall, and F1 score. The combination of optimal interval coefficients for aggregate sub-interval 1, aggregate sub-interval 2, and aggregate sub-interval 3 is 0.2, 0.3, and 0.5, respectively. The precision, recall, and F1 score for aggregate sub-interval 1 are 90%, 75%, and 82%, respectively; the precision, recall, and F1 score for aggregate sub-interval 2 are 95%, 92%, and 93%, respectively; and the precision, recall, and F1 score for aggregate sub-interval 3 are 92%, 90%, and 91%, respectively. If we perform single-objective optimization for precision, we can directly select aggregate sub-interval 2, which has the highest precision, as the optimal aggregate sub-interval. If we perform comprehensive optimization for both precision and F1 score, we calculate the combined score of these two scores, and aggregate sub-interval 2, which has both the highest precision and F1 score, can be selected as the optimal aggregate sub-interval.

[0088] S360. Input the actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value.

[0089] S370. Determine whether the electrical equipment under test has an arcing fault based on the comparison between the actual arcing probability value and the preset arcing probability threshold.

[0090] The technical solution of this embodiment, by taking all aggregated sub-intervals whose average energy spectral density satisfies the energy spectral density threshold as the optimal sub-interval set, can ensure complete coverage of arcing features. It can solve the technical problems of insufficient targeting and feature omission in the fixed frequency filtering in the prior art, and can accurately and stably select the optimal arcing feature interval. Furthermore, by using cross-validation to obtain the optimal aggregated sub-interval, the adaptability of model parameters and interval features can be optimized, achieving the technical effects of adaptive operation under complex working conditions, high-precision detection, low false alarms and false negatives, and strong stability.

[0091] In one embodiment, Figure 4 This is a flowchart of another arc fault detection method provided by an embodiment of the present invention. This embodiment further refines the process of determining the arc fault detection result based on the above embodiments. Figure 4 As shown, this embodiment includes:

[0092] S410. Based on the characteristics of the electrical signal, divide and aggregate the actual electrical signal of the entire range of the electrical equipment under test to obtain the aggregated sub-range actual electrical signal.

[0093] S420. Determine the optimal aggregation sub-interval based on the energy spectral density of the actual electrical signal in each aggregation sub-interval within a preset multi-frame electrical signal, and the energy spectral density threshold.

[0094] S430. Input the actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value.

[0095] S440. If the actual arcing probability value is less than the first actual arcing probability threshold, it is determined that the electrical equipment under test has an arcing fault.

[0096] S450. If the actual arcing probability value is greater than or equal to the first actual arcing probability threshold and less than the second actual arcing probability threshold, it is determined that the electrical equipment under test is suspected of having an arcing fault.

[0097] S460. If the actual arcing probability value is greater than or equal to the second actual arcing probability threshold, it is determined that the electrical equipment under test has not experienced an arcing fault.

[0098] S470. In the event of an arcing fault in the electrical equipment under test, generate a Level 1 alarm message and record the arcing fault information.

[0099] In one example, when an arcing fault occurs in the electrical equipment under test, a Level 1 alarm message can be generated and triggered, such as an audible and visual alarm or a remote notification; at the same time, fault information, such as the time of occurrence, location, and FS score, can be recorded; and emergency protection measures can be implemented, such as switching circuits and reducing power, to prevent the fault from spreading.

[0100] S480. In the event that the electrical equipment under test is suspected of having an arcing fault, generate a level 2 alarm message and activate the fault detection enhancement mode.

[0101] In one example, if the electrical equipment under test is suspected of having an arcing fault, a secondary alarm message can be generated and triggered to remind maintenance personnel to pay attention; and, an enhanced fault detection mode can be activated, such as increasing the signal acquisition frequency and strengthening feature analysis to continuously monitor status changes; and, emergency shutdown can be temporarily suspended to balance detection accuracy and equipment operational stability.

[0102] In one example, if the electrical equipment under test does not experience an arcing fault, an alarm may not be triggered, and the testing may continue at the normal frequency to maintain normal operation.

[0103] The technical solution of this embodiment generates alarm information of different levels when different types of faults occur in the electrical equipment under test. Emergency protection can be achieved by judging the fault level, and practicality can be balanced by judging the suspected level. This can achieve the effects of high-precision detection under complex working conditions, reducing false alarms and missed alarms, and ensuring the safe operation of equipment.

[0104] In one embodiment, before dividing and aggregating the actual electrical signals of the entire range of the electrical equipment under test according to the characteristics of the electrical signals to obtain the actual electrical signals of the aggregated sub-ranges, the method further includes: acquiring the original electrical signals of the entire range of the electrical equipment under test; and performing signal preprocessing operations on the original electrical signals of the entire range to obtain the actual electrical signals of the entire range.

[0105] In one example, the raw electrical signal across the entire range refers to the raw electrical signal across the entire range that has been acquired from the electrical equipment under test before any signal preprocessing has been performed. Signal preprocessing may include operations such as signal amplification and / or signal filtering. In this embodiment, the raw electrical signal across the entire range can be acquired, and preprocessing operations such as filtering and amplification can be performed on the raw electrical signal across the entire range to eliminate noise interference and obtain the actual electrical signal across the entire range.

[0106] The purpose of this invention is to provide a strategy for determining the arc detection filtering interval based on threshold clustering and spectral feature fusion. This strategy can dynamically adjust the detection parameters and strategy of the filtering interval to improve the accuracy of arc detection under complex working conditions and reduce the false alarm rate and false negative rate. Figure 5 This is a block diagram of an arc fault detection system provided in an embodiment of the present invention. Figure 5 As shown, the arc fault detection system provided in this embodiment may include: a signal acquisition unit, a signal amplifier, a signal filter, and a microprocessor; wherein, the signal acquisition unit is used to acquire current signals; the signal amplifier can also be called a signal amplification circuit; the signal filter is used for dynamic interval selection; and the microprocessor is used to execute the interval selection algorithm and train the arc detection model. The acquired full-range original electrical signals are first amplified, and the amplified signals are dynamically selected through model training and the interval selection algorithm to obtain the optimal aggregate sub-interval. Finally, the model is retrained until the trained model meets the requirements, and the arc fault determination is performed using the model that meets the requirements.

[0107] Figure 6 This is a schematic diagram of a frequency domain waveform with a relatively light load under a relatively large current, provided by an embodiment of the present invention; Figure 7 This is a schematic diagram of a frequency domain waveform with a relatively small current and a relatively light load, provided by an embodiment of the present invention; Figure 8 This is a schematic diagram of a frequency domain waveform with a relatively small current and a relatively heavy load, provided by an embodiment of the present invention. Figure 6 and 7For example, the frequency domain waveform with a relatively light load in the same frequency band, such as Figure 6 and 7 The diagram shows the frequency domain waveforms of larger and smaller currents under a relatively light load in the same frequency band, with the arcing frequency domain waveform and the non-arcing frequency domain waveform labeled respectively. Through this... Figure 6 and 7 It is readily apparent that the differences in frequency domain waveforms between arcing and non-arcging under different current conditions are evident. Under high current conditions, the distinction between arcing and non-arcging is more pronounced in the mid-frequency range, while under low current conditions, the distinction is more pronounced in the low-frequency and mid-frequency ranges. However, this is not static. The frequency bands exhibited by arcing and non-arcging also differ depending on the load type and current. This is primarily due to the varying electrical characteristics of different loads and the different interactions between arc characteristics and the load under different current conditions. In actual electrical equipment operation, most loads are multifaceted loads. Under different current conditions, the interaction between these characteristics and the arc leads to varying degrees of distinction between arcing and non-arcging in different frequency bands. Figure 8 The waveforms of a low-current load with a heavy load are shown, with arcing and non-arcing characteristics being more pronounced in the low-frequency range. Therefore, using dynamic selection of different frequency bands for filtering can effectively utilize these significantly different frequency band signals and filter out noise interference signals, reducing the occurrence of false positives and false negatives.

[0108] Figure 9 This is a flowchart provided by an embodiment of the present invention, illustrating the process from acquiring raw electrical signals across the entire circuit to detecting arcing faults. Figure 9 The diagram illustrates the overall process from acquiring the original electrical signals across the entire range to final arc detection. First, voltage or current signals across the entire range are acquired as the original electrical signals. Then, preprocessing operations such as filtering and amplification are performed to eliminate noise interference. A convolutional neural network model is used to train the preprocessed electrical signals across the entire range, obtaining an initial arc detection model. Next, the entire range is divided into N initial sub-ranges, and the arc probability value for each initial sub-range is obtained based on the initial arc detection model. A clustering algorithm is used to divide the initial sub-ranges into M (M≤N) aggregated sub-ranges. Then, combined with preset multiple frames of electrical signals (i.e., pre-acquired multiple frames of electrical signals), the optimal sub-range set is obtained based on the energy spectral density of each aggregated sub-range. The optimal aggregated sub-range is selected using interval coefficients, and the circuit parameters are adjusted to retrain the arc detection model. Finally, the optimized target arc detection model is used to analyze the real-time signal to determine if an arc phenomenon exists. If the determination result indicates the presence of an arc, an alarm is immediately triggered, and corresponding protective measures are taken, such as cutting off the circuit, further providing strong protection for the safe operation of electrical equipment.

[0109] Figure 10 This is a schematic diagram illustrating the implementation of initial sub-interval division and arc probability value according to an embodiment of the present invention, as shown below. Figure 10 As shown, a detailed process of initial sub - interval division and calculation of the arcing probability value for each initial sub - interval is described. According to the operating characteristics of electrical equipment and the signal change law, set interval division rules to divide into N initial sub - intervals. For example, if the signal change is relatively stable, it can be divided at a fixed time interval; if the signal change fluctuates greatly, it can be dynamically divided according to the change points of signal intensity (such as the reciprocal peak value); for complex working conditions, an adaptive time window is adopted to adjust the window length in real - time according to signal entropy. Input the pre - processed signal within each initial sub - interval into the initial detection model, and the initial arcing detection model outputs the arcing probability value corresponding to each initial sub - interval (that is , i = 1, 2, …, N).

[0110] Figure 11 is a schematic diagram of the implementation of aggregating sub - interval division and determining the optimal aggregating sub - intervals provided by an embodiment of the present invention. As Figure 11 shown, it shows the aggregating sub - interval division and determination of the optimal sub - interval set. According to the arcing probability values of the initial sub - intervals, use a clustering algorithm (K - means clustering or threshold clustering), such as the adaptive threshold clustering method to merge N initial sub - intervals into M aggregating sub - intervals (M is a positive integer and M ≤ N). The specific process is as follows: calculate the average value of the arcing probability of all initial sub - intervals , set the clustering threshold as the first aggregating arcing probability threshold T1 and the second aggregating arcing probability threshold T2 based on the average value of the arcing probability; classify the sub - intervals with < T1 into the "low - probability value interval set", those with T1 ≤ ≤ T2 into the "medium - probability value interval set", and those with > T2 into the "high - probability value interval set"; merge the initial sub - intervals within each interval set according to the continuous - order merging principle into M aggregating sub - intervals.

[0111] Figure 12 is a schematic diagram of the display of the energy spectral density of multiple frames of electrical signals provided by an embodiment of the present invention. Collect multiple frames (such as the number of frames K = 5 - 20, and the duration of each frame signal is the same as that of the full - interval signal) of continuous pre - processed signals, and calculate the energy spectral density ESD (as Figure 12 shown) of each aggregating sub - interval in the K - frame signals: perform Fourier transform on the signal within the aggregating sub - interval to obtain the frequency - domain signal X(f) (f is the frequency); the energy spectral density calculation formula is: , where T is the duration of the aggregating sub - interval; calculate the average value of the ESD of each aggregating sub - interval in the K - frame signals. The larger the average value of this ESD, the more concentrated the energy of the signal within this interval, and the more significant the arcing feature; set the energy spectral density threshold , and for the average value The aggregated sub-intervals are included in the optimal sub-interval set to ensure that the aggregated sub-intervals in the optimal sub-interval set can completely cover the high-energy characteristic region of the arc signal.

[0112] Figure 13 This is a flowchart illustrating the implementation of model optimization and arc detection protection provided by an embodiment of the present invention, as shown below. Figure 13 The diagram illustrates the flowchart for model optimization and arc detection protection. Each aggregated sub-interval in the optimal sub-interval set is assigned an interval coefficient Kj, where j represents the number of aggregated sub-intervals within the optimal sub-interval set, and Kj is a weighting factor reflecting the importance of the interval. First, the coefficients of each aggregated sub-interval within the optimal sub-interval set are initialized. A portion of the training data is selected as a validation set to evaluate the effectiveness of different coefficient combinations. Different interval coefficients are applied to each aggregated sub-interval, and each combination is applied to the model using cross-validation and evaluated on the validation set. By comparing the detection performance under different coefficient combinations, the optimal interval coefficient combination is selected. Evaluation metrics for detection performance may include accuracy, recall, and F1 score. Finally, using the optimized interval coefficients, the interval range that achieves the best detection performance is selected. Circuit parameters are adjusted in conjunction with key electrical equipment indicators to make the signal characteristics within the optimal interval more significant. The model is then retrained to obtain the final arc detection model, further improving the accuracy and reliability of arc detection. The system reads the electrical signals collected in real time and outputs the arcing probability value of the real-time signal through the final detection model. This arcing probability value is compared with the pre-configured first actual arcing probability threshold T3 and second actual arcing probability threshold T4, and the final arcing judgment is based on the number of judgments optimized. If FS < T3, an arcing fault is determined, the system immediately triggers a level one alarm, and records various information about the arcing occurrence, such as time and location; at the same time, corresponding protection measures are taken, such as cutting off the circuit or reducing power; if T3 ≤ FS < T4, it is determined to be a suspected arcing fault, triggers a level two alarm, and starts the enhanced detection mode; if FS ≥ T4, it is determined to be a normal state, and normal detection continues.

[0113] Figure 14 This is a schematic diagram illustrating the false alarms and false negatives of arc detection using a traditional interval model, as provided in an embodiment of the present invention. Figure 15 This is a schematic diagram illustrating the false alarms and false negatives of arc detection using an optimized target arc detection model, as provided in an embodiment of the present invention.

[0114] Figure 14 To address the false alarms and missed detection issues associated with using the traditional interval model for arc detection, from... Figure 14 As can be seen, the false alarm and false negative scores fluctuate greatly, indicating that the model has a certain degree of false alarm and false negative. Figure 15To address the false alarms and false negatives in arc detection using the optimized target arc detection model of this invention, compared to... Figure 14 The fluctuations in false alarm and false negative scores were significantly reduced, and the false alarm and false negative situations were improved. This indicates that the invention has improved the accuracy of detection to a certain extent and reduced the false alarm and false negative rates.

[0115] In one embodiment, Figure 16 This is a structural schematic diagram of an arc fault detection device provided in an embodiment of the present invention. Figure 16 As shown, the arcing fault detection device includes: a signal acquisition unit 20, a signal processor 21, and a microprocessor 11; the output terminal of the signal acquisition unit 20 is connected to the input terminal of the signal processor 21, and the output terminal of the signal processor is connected to the microprocessor 11.

[0116] The signal acquisition unit 20 is used to acquire the original full-range actual electrical signal of the electrical equipment under test, and sends the original full-range actual electrical signal to the signal processor 21. The signal processor 21 performs filtering and amplification operations on the original full-range actual electrical signal to obtain the actual full-range actual electrical signal. The microprocessor 11 divides and aggregates the full-range actual electrical signal of the electrical equipment under test according to the characteristics of the electrical signal to obtain aggregated sub-interval actual electrical signals. The optimal aggregated sub-interval is determined according to the energy spectral density of each aggregated sub-interval actual electrical signal in the preset multi-frame electrical signal and the energy spectral density threshold. The aggregated sub-interval actual electrical signal corresponding to the optimal aggregated sub-interval is input into the target arc detection model that matches the optimal aggregated sub-interval to obtain the actual arc probability value. The actual arc probability value is compared with the preset arc probability threshold to determine whether the electrical equipment under test has an arc fault.

[0117] The arc fault detection device provided in the embodiments of the present invention can execute the arc fault detection method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.

[0118] Figure 17 This is a schematic diagram of another arc fault detection device provided in an embodiment of the present invention. Figure 17As shown, the arc fault detection device, in addition to including a signal acquisition unit, a signal processor, and a microprocessor, may also include a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, which is communicatively connected to at least one microprocessor 11. The memory stores computer programs executable by at least one microprocessor. The microprocessor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the arc fault detection device 10. The microprocessor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0119] Multiple components in the arc fault detection device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, optical disk, etc.; and a communication unit 19, such as a network card, modem, wireless transceiver, etc. The communication unit 19 allows the arc fault detection device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0120] Microprocessor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of microprocessor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any suitable processor, controller, microcontroller, etc. Microprocessor 11 performs the various methods and processes described above, such as arc fault detection methods.

[0121] In some embodiments, the arcing fault detection method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the arcing fault detection device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by microprocessor 11, one or more steps of the arcing fault detection method described above may be performed. Alternatively, in other embodiments, microprocessor 11 may be configured to perform the arcing fault detection method by any other suitable means (e.g., by means of firmware).

[0122] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0123] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0124] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0125] To provide user interaction, the systems and techniques described herein can be implemented on an arc fault detection device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the arc fault detection device. Other types of devices can also be used to provide user interaction; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0126] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0127] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0128] This invention also provides a computer program product, including a computer program that, when executed by a processor, can implement the arcing fault detection method provided in any embodiment of this application.

[0129] In the implementation of the computer program product, computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0130] This invention also provides an arcing fault detection system, which includes: an electrical device under test and an arcing fault detection device as described in any embodiment of this invention.

[0131] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0132] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for detecting arcing faults, characterized in that, The method includes: Based on the characteristics of electrical signals, the actual electrical signals of the entire range of the electrical equipment under test are divided and aggregated to obtain the aggregated sub-range actual electrical signals. The optimal aggregation sub-interval is determined based on the energy spectral density of the actual electrical signal in each aggregation sub-interval within a preset multi-frame electrical signal, and the energy spectral density threshold. The actual electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval is input into the target arc detection model that matches the optimal aggregate sub-interval to obtain the actual arc probability value; The presence of an arcing fault in the electrical equipment under test is determined based on the comparison between the actual arcing probability value and the preset arcing probability threshold. The step of dividing and aggregating the actual electrical signals of the entire range of the electrical equipment under test according to the characteristics of the electrical signals to obtain the aggregated sub-range actual electrical signals includes: If the electrical signal changes smoothly, the actual electrical signal of the entire range of the electrical equipment under test is divided into intervals according to fixed time intervals to obtain multiple initial sub-interval actual electrical signals. If there are sudden peaks in the electrical signal, the actual electrical signal of the entire range of the electrical equipment under test is dynamically divided into multiple initial sub-range actual electrical signals based on the change points of the electrical signal strength. If the electrical signal does not have a fixed pattern of change, an adaptive time window division method is used to dynamically divide the actual electrical signal of the entire range of the electrical equipment under test into multiple initial sub-range actual electrical signals. The actual electrical signal of each initial sub-interval is input into the initial arc detection model to obtain the arc probability value of the corresponding initial sub-interval actual electrical signal; The average arcing probability is determined based on the arcing probability value of the actual electrical signal in each initial sub-interval and the number of initial sub-intervals; The first aggregate arcing probability threshold and the second aggregate arcing probability threshold are determined according to the average arcing probability; wherein, the first aggregate arcing probability threshold is less than the second aggregate arcing probability threshold; The actual electrical signals of the multiple initial sub-intervals are aggregated based on the first aggregated arcing probability threshold and the second aggregated arcing probability threshold to obtain multiple aggregated sub-interval actual electrical signals.

2. The method according to claim 1, characterized in that, The step of determining the optimal aggregation sub-interval based on the energy spectral density of the actual electrical signal in each aggregation sub-interval within a preset multi-frame electrical signal, and an energy spectral density threshold, includes: Determine the energy spectral density of the actual electrical signal in each of the said aggregate sub-intervals in the pre-acquired multi-frame electrical signals; All aggregated sub-intervals whose average energy spectral density is greater than or equal to the energy spectral density threshold are taken as the optimal sub-interval set; The optimal interval coefficients corresponding to each aggregated subinterval in the optimal subinterval set are determined by cross-validation. Using the optimal interval coefficient, the aggregated sub-interval with the best detection effect is selected from the set of optimal sub-intervals as the optimal aggregated sub-interval.

3. The method according to claim 2, characterized in that, Determining the energy spectral density of the actual electrical signal in each of the aggregated sub-intervals within the pre-acquired multi-frame electrical signals includes: Perform a Fourier transform on the actual electrical signal of each of the said aggregate sub-intervals to obtain the frequency domain signal; The energy spectral density of the corresponding aggregate sub-interval is determined by the ratio between the square of the frequency domain signal and the duration of the corresponding aggregate sub-interval.

4. The method according to claim 1, characterized in that, Determining whether the electrical equipment under test has an arcing fault based on the comparison between the actual arcing probability value and the preset arcing probability threshold includes: If the actual arcing probability value is less than the first actual arcing probability threshold, it is determined that the electrical equipment under test has an arcing fault. If the actual arcing probability value is greater than or equal to the first actual arcing probability threshold and less than the second actual arcing probability threshold, it is determined that the electrical equipment under test is suspected of having an arcing fault. If the actual arcing probability value is greater than or equal to the second actual arcing probability threshold, it is determined that the electrical equipment under test has not experienced an arcing fault.

5. The method according to claim 4, characterized in that, After determining whether the electrical equipment under test has an arcing fault based on the comparison result between the actual arcing probability value and the preset arcing probability threshold, the method further includes: In the event of an arcing fault in the electrical equipment under test, a Level 1 alarm message is generated and the arcing fault information is recorded. If the electrical equipment under test is suspected of having an arcing fault, a level 2 alarm message is generated and the fault detection enhancement mode is activated.

6. The method according to any one of claims 1-5, characterized in that, Before dividing and aggregating the actual electrical signals of the entire range of the electrical equipment under test according to the electrical signal characteristics to obtain the aggregated sub-range actual electrical signals, the method further includes: Acquire the raw electrical signals of the electrical equipment under test across the entire range; The original electrical signals of the entire range are preprocessed to obtain the actual electrical signals of the entire range.

7. The method according to any one of claims 1-5, characterized in that, The construction process of the target arc detection model includes: Based on the characteristics of electrical signals, the full-range training electrical signals in the electrical signal training set are divided into intervals and aggregated to obtain aggregated sub-interval training electrical signals; The optimal aggregation sub-interval is determined based on the energy spectral density of the training electrical signal in a preset multi-frame electrical signal for each aggregation sub-interval, and the energy spectral density threshold. The initial arc detection model is trained using the training electrical signal of the aggregate sub-interval corresponding to the optimal aggregate sub-interval, and the target arc detection model is obtained.

8. An arc fault detection device, characterized in that, The arcing fault detection device includes: a signal acquisition unit, a signal processor, and a microprocessor; the output terminal of the signal acquisition unit is connected to the input terminal of the signal processor, and the output terminal of the signal processor is connected to the microprocessor. The signal acquisition unit is used to acquire the original full-range actual electrical signal of the electrical equipment under test, send the original full-range actual electrical signal to the signal processor, and the signal processor performs filtering and amplification operations on the original full-range actual electrical signal to obtain the actual full-range actual electrical signal. The microprocessor divides and aggregates the full-range actual electrical signal of the electrical equipment under test according to the characteristics of the electrical signal to obtain aggregated sub-interval actual electrical signals. The optimal aggregated sub-interval is determined according to the energy spectral density of each aggregated sub-interval actual electrical signal in a preset multi-frame electrical signal and the energy spectral density threshold. The actual electrical signal of the aggregated sub-interval corresponding to the optimal aggregated sub-interval is input into the target arc detection model that matches the optimal aggregated sub-interval to obtain the actual arc probability value. The actual arc probability value and the preset arc probability threshold are compared to determine whether the electrical equipment under test has an arc fault. The step of dividing and aggregating the actual electrical signals of the entire range of the electrical equipment under test according to the characteristics of the electrical signals to obtain the aggregated sub-range actual electrical signals includes: If the electrical signal changes smoothly, the actual electrical signal of the entire range of the electrical equipment under test is divided into intervals according to fixed time intervals to obtain multiple initial sub-interval actual electrical signals. If there are sudden peaks in the electrical signal, the actual electrical signal of the entire range of the electrical equipment under test is dynamically divided into multiple initial sub-range actual electrical signals based on the change points of the electrical signal strength. If the electrical signal does not have a fixed pattern of change, an adaptive time window division method is used to dynamically divide the actual electrical signal of the entire range of the electrical equipment under test into multiple initial sub-range actual electrical signals. The actual electrical signal of each initial sub-interval is input into the initial arc detection model to obtain the arc probability value of the corresponding initial sub-interval actual electrical signal; The average arcing probability is determined based on the arcing probability value of the actual electrical signal in each initial sub-interval and the number of initial sub-intervals; The first aggregate arcing probability threshold and the second aggregate arcing probability threshold are determined according to the average arcing probability; wherein, the first aggregate arcing probability threshold is less than the second aggregate arcing probability threshold; The actual electrical signals of the multiple initial sub-intervals are aggregated based on the first aggregated arcing probability threshold and the second aggregated arcing probability threshold to obtain multiple aggregated sub-interval actual electrical signals.

9. An arc fault detection system, characterized in that, The arcing fault detection system includes: the electrical equipment under test and the arcing fault detection device as described in claim 8.

Images