Arc fault detection method and related device

Abstract

The invention discloses an arc fault detection method and a related device. The method comprises the following steps: determining a first statistical characteristic of a current signal in a time window; performing waveform abrupt change analysis on the current signal to obtain a change interval with the maximum fluctuation amplitude in the current signal; extracting at least one time domain feature, at least one frequency domain feature and a plurality of second statistical features of the current signal; determining whether a transient event exists or not according to the first statistical characteristic, the real-time current data and the change interval; if the transient event exists, determining a target fault detection model according to the at least one time domain feature and the change interval; performing feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features to obtain a target feature; and determining a fault detection result according to the target feature and the target fault detection model. According to the invention, the fault detection efficiency and accuracy of the system can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system safety, and in particular to an arc fault detection method and related device. BACKGROUND

[0002] An electrical device will generate an arc in a circuit during actual operation. The arc is the most intense self-sustaining discharge phenomenon in gas discharge, and has characteristics of very high temperature, very small current, and short duration. The arc includes normal arc and fault arc. The fault arc can produce high temperature and high energy, and can ignite flammable and explosive materials around, causing fire and even explosion. Therefore, it is necessary to detect and clear the arc fault in time and accurately before the arc causes fire. However, the traditional fault arc detection technology is mainly based on the time-frequency characteristics of current and voltage. When facing complex loads and harmonic interference, the traditional method often shows limitations, and is prone to false positives and false negatives, resulting in low fault arc detection accuracy and poor reliability. SUMMARY

[0003] Embodiments of the present application provide an arc fault detection method and related device to improve the efficiency and accuracy of system fault detection.

[0004] In a first aspect, embodiments of the present application provide an arc fault detection method, comprising: determining a first statistical feature of a current signal in a time window, the current signal comprising an even number of sampling points, and the first statistical feature being used to represent a local offset level and an overall dispersion degree of the current signal; performing waveform mutation analysis on the current signal to obtain a change interval with the largest fluctuation amplitude in the current signal; extracting at least one time domain feature, at least one frequency domain feature, and a plurality of second statistical features of the current signal, the plurality of second statistical features being used to represent fluctuation amplitudes and morphological deviations of the current signal; determining whether a transient event exists according to the first statistical feature, real-time current data, and the change interval; if the transient event exists, determining a target fault detection model according to the at least one time domain feature and the change interval; performing feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; determining a fault detection result according to the target feature and the target fault detection model.

[0005] The determination of the first statistical feature of the current signal in the time window comprises: determining a variance of the current signal in the time window; splitting the time window to obtain a first sub-time window and a second sub-time window, a first number of sampling points being equal to a second number of sampling points, the first number of sampling points being a number of sampling points of the current signal in the first sub-time window, and the second number of sampling points being a number of sampling points of the current signal in the second sub-time window; determining a first mean value of the current signal in the first sub-time window, and determining a second mean value of the current signal in the second sub-time window; obtaining the first statistical feature according to the first mean value, the second mean value, and the variance.

[0006] wherein, the determining whether the transient event exists according to the first statistical feature, the real-time current data, and the change interval comprises: determining a change amount of the current signal according to the first mean value, the second mean value, the variance, and the real-time current data; detecting that the change amount is greater than a preset change amount, and determining an interval duration of the change interval; determining whether the transient event exists according to the interval duration.

[0007] wherein, the determining the change amount of the current signal according to the first mean value, the second mean value, the variance, and the real-time current data comprises: determining a first difference value of the first mean value and the second mean value, the first difference value being a positive number; determining a proportional value of the first difference value and the variance; performing mean value operation on the first mean value and the second mean value to obtain a third mean value; determining a second difference value of the real-time current data and the third mean value, the second difference value being a positive number; determining a product of the second difference value and the proportional value to obtain the change amount.

[0008] wherein, the performing feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature comprises: determining a first load type corresponding to the current signal; obtaining a monitoring requirement of the first load type; creating a first feature subset, the first feature subset at least including any two different dimension features in the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features; performing feature expansion processing on the first feature subset to obtain a second feature subset, the feature expansion processing being used to expand a single other feature into the current feature subset, the other feature being a feature other than the first feature subset from among the at least one time-domain feature, the at least one frequency-domain feature, and the plurality of second statistical features; determining a first correlation degree of the second feature subset with the monitoring requirement; determining a feature processing manner of the second feature subset according to a third difference value between the first correlation degree and a preset correlation degree, the feature processing manner including the feature expansion processing and / or feature elimination processing, the feature elimination processing being used to eliminate a single feature in the current feature subset; performing feature processing on the second feature subset according to the feature processing manner to obtain a third feature subset; determining a second correlation degree of the third feature subset with the monitoring requirement; detecting that a number of features in the third feature subset is a preset number, the second correlation degree is greater than or equal to the preset correlation degree, and each of the other features is processed by the feature expansion processing and / or the feature elimination processing, and then performing fusion processing on each feature in the third feature subset to obtain the target feature.

[0009] The determining of the feature processing manner of the second feature subset according to the third difference value between the first correlation degree and the preset correlation degree includes: if the third difference value is within a first preset range, the feature processing manner is the feature expansion processing; if the third difference value is within a second preset range, the feature processing manner is the feature elimination processing, an upper limit value of the first preset range being less than a lower limit value of the second preset range; if the third difference value is within a third preset range, the feature processing manner is the feature elimination processing and the feature expansion processing, the upper limit value of the second preset range being less than a lower limit value of the third preset range.

[0010] The method further includes: if the transient event does not exist, obtaining a second load type identified by a pre-transient event trigger; determining a target fault detection model according to the second load type; performing feature screening processing and feature fusion processing on the at least one time-domain feature, the at least one frequency-domain feature, and the plurality of second statistical features to obtain a target feature; determining a fault detection result according to the target feature and the target fault detection model.

[0011] In a second aspect, an arc fault detection apparatus is provided, comprising: a first determining unit configured to determine a first statistical feature of a current signal in a time window, the current signal comprising an even number of sampling points, the first statistical feature being used to represent a local offset level and an overall dispersion degree of the current signal; an analyzing unit configured to perform waveform mutation analysis on the current signal to obtain a variation interval with a maximum fluctuation amplitude in the current signal; an extracting unit configured to extract at least one time domain feature, at least one frequency domain feature and a plurality of second statistical features of the current signal, the plurality of second statistical features being used to represent a fluctuation amplitude and a shape deviation of the current signal; a second determining unit configured to determine whether a transient event exists according to the first statistical feature, real-time current data and the variation interval; a matching unit configured to determine a target fault detection model according to the at least one time domain feature and the variation interval if the transient event exists; a feature processing unit configured to perform feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features to obtain a target feature; a third determining unit configured to determine a fault detection result according to the target feature and the target fault detection model.

[0012] In a third aspect, an electronic device is provided, comprising a memory, a processor and executable program code stored in the memory and executable on the processor, the processor executes the executable program code to perform the steps of the method in the first aspect.

[0013] In a fourth aspect, a computer readable storage medium is provided, the computer readable storage medium stores executable program code, the executable program code comprises execution instructions for performing the steps of the method in the first aspect.

[0014] In a fifth aspect, a computer program product is provided, the computer program product comprises a non-transitory computer readable storage medium storing a computer program, the computer program is operable to cause a computer to perform some or all of the steps described in the first aspect of the embodiments. The computer program product can be a software installation package.

[0015] It can be seen that, in the embodiment of the present application, first, the first statistical feature of the current signal in the time window is determined, the current signal includes an even number of sampling points, and the first statistical feature is used to represent the local offset level and the overall dispersion degree of the current signal; then, waveform mutation analysis is performed on the current signal to obtain the change interval with the largest fluctuation amplitude in the current signal; then, at least one time domain feature, at least one frequency domain feature and a plurality of second statistical features of the current signal are extracted, and the plurality of second statistical features are used to represent the fluctuation amplitude and the morphological deviation of the current signal; then, whether there is a transient event is determined according to the first statistical feature, the real-time current data and the change interval; if the transient event exists, a target fault detection model is determined according to the at least one time domain feature and the change interval; then, the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features are subjected to feature screening processing and feature fusion processing to obtain a target feature; finally, a fault detection result is determined according to the target feature and the target fault detection model.

[0016] The present application determines the transient event by multi-dimensional data, which is beneficial to eliminate false events and improve the accuracy of determination; then, the matching of the fault detection model is triggered according to the transient event; and when the feature processing is performed, the time domain, frequency domain and statistical features are screened and combined to eliminate redundant features; then, the target feature obtained after the feature processing and the target fault detection model obtained by matching are combined for fault detection, which can make the fault detection result more in line with the actual situation of the load, thereby improving the pertinence and accuracy of fault detection, avoiding the misjudgment or omission problem caused by the mismatch of load characteristics, and improving the efficiency and accuracy of power system fault detection. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating labor.

[0018] Figure 1 is a system architecture diagram of a fault detection system provided by an embodiment of the present application; Figure 2 is a flowchart of a first arc fault detection method provided by an embodiment of the present application; Figure 3 is a flowchart of determining whether there is a transient event provided by an embodiment of the present application; Figure 4 is a flowchart of a second arc fault detection method provided by an embodiment of the present application; Figure 5 is a flowchart of a third arc fault detection method provided by an embodiment of the present application; Figure 6 is a functional unit composition block diagram of an arc fault detection device provided by an embodiment of the present application; Figure 7 is a functional unit composition block diagram of another arc fault detection device provided by an embodiment of the present application; Figure 8 is a structural diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative work fall within the scope of protection of the present application.

[0020] The terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish different objects, and are not used to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but can optionally include other steps or units not listed or can optionally include other steps or units inherent to the process, method, product or device.

[0021] In this document, the term "embodiment" means that the specific features, structures or characteristics described in connection with the embodiment can be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily mean the same embodiment, nor is it independent or alternative to other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0022] Arcs are generated in the circuit during the actual operation of electrical equipment. Arc is the most intense self-sustaining discharge phenomenon in gas discharge, which has the characteristics of high temperature, small current and short duration. Among them, arc includes normal arc and fault arc. Fault arc can produce high temperature and high energy, which can ignite the surrounding flammable and explosive materials, causing fire and even explosion. Therefore, it is necessary to detect and clear the arc fault in time and accurately before the arc causes fire. However, the traditional fault arc detection technology is mainly based on the time-frequency characteristics of current and voltage. When facing complex load and harmonic interference, the traditional method often shows limitations, and false positives and false negatives are easy to occur, resulting in low fault arc detection accuracy and poor reliability.

[0023] In addition, arc fault signals are usually accompanied by noise interference, and traditional methods perform poorly in low signal-to-noise ratio environments, affecting the accuracy of arc fault detection.

[0024] To solve at least one of the above problems, the embodiments of the present application provide an arc fault detection method and related device, which will be described in detail below with reference to the accompanying drawings.

[0025] Please refer to Figure 1 , Figure 1 is a system architecture diagram of a fault detection system provided by the embodiments of the present application. As shown in Figure 1 , the fault detection system 100 includes a discrimination module 101, a feature extraction module 102 and a fault detection module 103, and the discrimination module 101, the feature extraction module 102 and the fault detection module 103 are connected in communication with each other.

[0026] Among them, the discrimination module 101 is used to discriminate whether a transient event occurs in a current sliding window, analyzes the current signal in the sliding window segment by segment, determines the mutation or abnormal change information in the current signal by calculating the mean and variance of the current signal in each window, and deletes false events according to the duration of abnormal change to ensure that false positives are eliminated, and finally determines whether a transient event occurs.

[0027] Among them, the feature extraction module 102 is used to extract the time domain features, frequency domain features and statistical features of the current signal, and according to the load type determined by the fault detection module 103, matches the corresponding feature evaluation index, and according to the feature evaluation index, the extracted features are subjected to feature screening processing and feature fusion processing to obtain target features.

[0028] The fault detection module 103 is configured to receive the information sent by the discrimination module 101 and the feature extraction module 102, trigger the identification of the load type under the condition that the transient event occurs, determine the load type according to the duration of the abnormal change, the time range where the abnormal change occurs and the transient feature, determine the corresponding fault detection model according to the load type, input the target feature into the fault detection model, and output the fault detection result. Under the condition that the transient event does not occur, the pre-load type triggered by the previous transient event is determined as the current load type, and the fault detection is performed according to the fault detection model corresponding to the load type.

[0029] Based on this, the present application provides an arc fault detection method and related device, which will be described in detail below with reference to the accompanying drawings.

[0030] Please refer to Figure 2 , Figure 2 is a flowchart of an arc fault detection method provided by an embodiment of the present application, as shown in Figure 2 , the method comprises the following steps: S210, determining a first statistical feature of the current signal in the time window.

[0031] The current signal comprises an even number of sampling points, and the first statistical feature is used to represent the local offset level and the overall dispersion degree of the current signal.

[0032] The system captures the voltage and current signals in real time, and the time window is sliding, which is advanced by a preset step each time, for example, 1 step each time.

[0033] The current signal refers to the continuous or discrete waveform data of the entire current changing over time.

[0034] Each time window comprises 2n current sampling points, the current signal in the time window is analyzed segment by segment, the mean value of the current signal in each segment window is calculated, and the variance of the current signal in the entire window is calculated to obtain the current current signal sample in the time window; the transient change of the current waveform is determined according to the current current signal sample, the mean value and the variance.

[0035] The current current signal sample is the current value in real-time monitoring, which is the latest data point in the sliding window.

[0036] In a possible embodiment, the determining the first statistical feature of the current signal in the time window comprises: determining a variance of the current signal in the time window; splitting the time window to obtain a first sub-time window and a second sub-time window, the first sub-time window and the second sub-time window have the same number of sampling points, the number of sampling points in the first sub-time window is the number of sampling points of the current signal in the first sub-time window, and the number of sampling points in the second sub-time window is the number of sampling points of the current signal in the second sub-time window; determining a first mean value of the current signal in the first sub-time window; and determining a second mean value of the current signal in the second sub-time window; and obtaining the first statistical feature according to the first mean value, the second mean value, and the variance.

[0037] The first statistical feature comprises a mean value corresponding to each sub-time window and a variance corresponding to the time window.

[0038] The time window can be equally divided into two parts, each sub-time window comprises n current sampling points, and the mean value of the current signal in each sub-time window is calculated, that is, the mean value of n current sampling points in the first half part and the mean value of n current sampling points in the second half part are calculated, and the local offset level of the current signal is reflected by the first mean value and the second mean value.

[0039] The variance of 2n current sampling points is calculated at the same time, and the overall dispersion degree of the current signal is reflected by the variance.

[0040] According to the mean value corresponding to each sub-time window and the variance corresponding to the time window, the mutation or abnormal change in the current signal can be detected.

[0041] In a possible embodiment, the length of the time window can be adjusted according to the variation characteristics of the actual current signal, to balance the detection accuracy and the calculation efficiency.

[0042] In a possible embodiment, an automatic adaptive algorithm can be used to analyze the actual operating conditions of the power grid in real time, and then the sampling frequency and the sampling period can be dynamically adjusted to improve the detection accuracy and reduce the calculation burden.

[0043] Specifically, the sampling frequency can be dynamically adjusted according to the power grid load and the arc fault characteristics. When the power grid load is low and the current changes smoothly, the sampling frequency can be reduced; when the power grid load is high or the probability of arc fault occurrence increases, the sampling frequency can be increased to ensure that the high-frequency components of the arc fault are captured in time.

[0044] Specifically, the sampling period can be dynamically adjusted according to the current waveform change rate. When the current fluctuation amplitude is large, the sampling period is shortened to improve real-time performance; when the current fluctuation amplitude is small, the sampling period is lengthened to reduce data redundancy, reduce storage and calculation pressure, and avoid interference of invalid data on the analysis process.

[0045] In one possible embodiment, the number of sampling points is dynamically adjusted according to the adjustment of the time window length, and / or the adjustment of the sampling frequency, and / or the adjustment of the sampling period.

[0046] Among them, according to the adjusted number of sampling points, the sub-time window is re-divided.

[0047] For example, a first threshold and a second threshold are preset, the second threshold is greater than the first threshold, if the number of added sampling points is greater than the first threshold, when splitting the time window, one sub-time window is added, that is, the time window is evenly split into three sub-time windows; if the number of added sampling points is greater than the second threshold, when splitting the time window, two sub-time windows are added, that is, the time window is evenly split into four sub-time windows.

[0048] S220, waveform mutation analysis is performed on the current signal to obtain a change interval with the largest fluctuation amplitude in the current signal.

[0049] Among them, the time period when the waveform of the current signal changes dramatically is identified, and the interval with the most significant fluctuation amplitude is screened out to obtain the change interval.

[0050] Among them, the current signal is subjected to data preprocessing such as filtering and normalization, and mutation feature extraction such as instantaneous change amount and change rate, and the mutation strength is determined according to the mutation feature. Among them, the instantaneous change amount refers to the current difference between adjacent sampling points, and the larger the difference, the more obvious the mutation; the change rate refers to the current change amount per unit time, which reflects the speed of mutation.

[0051] Among them, the continuous current signal is divided into several intervals through a sliding window, the mutation strength in each interval is calculated, and finally the interval with the highest mutation strength, i.e. the change interval with the largest fluctuation amplitude, is screened out.

[0052] S230, at least one time domain feature, at least one frequency domain feature and a plurality of second statistical features of the current signal are extracted.

[0053] Among them, the plurality of second statistical features are used to represent the fluctuation amplitude and the shape deviation of the current signal.

[0054] Among them, the time domain feature is a feature based on the time dimension, which is extracted from the time sequence of the current signal and is used to reflect the instantaneous change or overall trend of the signal with time. It can include peak value, mean value, root mean square value, kurtosis, etc.

[0055] The frequency domain feature is a feature based on a frequency dimension, and the time domain signal is converted to a frequency domain by a Fourier transform, a wavelet transform or the like, to reflect frequency components and energy distribution contained in the signal.

[0056] When the arc fault occurs, specific high-frequency noise components appear in the frequency domain, and the high-frequency components in the arc fault signal can be effectively extracted by converting the current signal from the time domain to the frequency domain through a fast Fourier transform.

[0057] Specifically, the frequency domain feature can include a main frequency, i.e., a frequency component with the highest energy proportion in the signal; can include a harmonic amplitude, i.e., an amplitude of a high-order harmonic other than the main frequency, to reflect the degree of spectral pollution of the signal; can include a spectral entropy to measure the uniformity of the distribution of the frequency components, a high spectral entropy indicating that the signal contains multiple frequencies, and a low entropy indicating that the frequency components are single, such as a stable sine wave; and can include a bandwidth to represent the distribution range of the main frequency components of the signal.

[0058] The second statistical feature can include a variance, a standard deviation, a range, a skewness, a peak factor, etc. The variance, the standard deviation and the range are used to reflect the fluctuation amplitude of the current signal, and the skewness and the peak factor are used to describe the distribution characteristics and the morphological deviation of the current signal.

[0059] The time domain feature, the frequency domain feature and the second statistical feature are combined to distinguish the fault signal and the normal signal.

[0060] S240, determining whether a transient event exists according to the first statistical feature, real-time current data and the change interval.

[0061] In one possible embodiment, referring to Figure 3 , Figure 3 is a flowchart for determining whether a transient event exists, provided by an embodiment of the present application, Figure 3 shows a step of determining whether a transient event exists according to the first statistical feature, real-time current data and the change interval, and specifically includes the following steps: S310, determining a change amount of the current signal according to the first mean value, the second mean value, the variance and the real-time current data.

[0062] The real-time current data is a current value in real-time monitoring, and is the latest data in the sliding window. According to the mean value, the variance and the real-time current data of the current signal, a sudden change or abnormal change in the current signal is detected.

[0063] In a possible embodiment, the determining the change amount of the current signal according to the first mean value, the second mean value, the variance and the real-time current data comprises: determining a first difference value of the first mean value and the second mean value, the first difference value being a positive number; determining a proportional value of the first difference value and the variance; performing mean value operation on the first mean value and the second mean value to obtain a third mean value; determining a second difference value of the real-time current data and the third mean value, the second difference value being a positive number; and determining a product of the second difference value and the proportional value to obtain the change amount.

[0064] wherein the change amount of the current signal is as follows: , wherein, is the first mean value, is the second mean value, is the variance, and the real-time current data.

[0065] In a possible embodiment, if the time window is divided into three sub-time windows, the mean value of each sub-time window is calculated to obtain the first mean value, the second mean value and the third mean value, and the variance of the whole time window is calculated. Then, the difference value between each pair of mean values is calculated to obtain the first difference value, the second difference value and the third difference value, the above difference values being positive numbers, the mean value of the three difference values is calculated to obtain a fourth mean value, the proportional value of the fourth mean value and the variance is calculated, and the mean value of the first mean value, the second mean value and the third mean value is calculated to obtain a fifth mean value. The difference value between the real-time current data and the fifth mean value is calculated to obtain a fourth difference value, the fourth difference value being a positive number, and the product of the fourth difference value and the proportional value is calculated to obtain the change amount of the current signal.

[0066] S320, detecting that the change amount is greater than a preset change amount, and determining the interval duration of the change interval.

[0067] wherein, when the change amount of the current signal exceeds a set threshold value, whether it is a false event is determined according to the interval duration of the change interval, an event with a too short interval duration is determined as a false event, and is removed.

[0068] wherein, the false event refers to a signal that is misjudged as a transient event, and actually has no corresponding physical process, and may be caused by measurement noise, equipment error, etc. The duration of the false event may be extremely short or irregular.

[0069] S330, determining whether there is a transient event according to the interval duration.

[0070] The event with an interval duration less than the minimum possible time of the physical process can be determined as a false event; the event with an interval duration inconsistent with the physical process can be determined as a false event, for example, an upper limit is set according to the typical duration of a specific transient type, and if the interval duration exceeds the upper limit, the event is determined as a false event; and the event with irregular duration can also be determined as a false event.

[0071] The transient event refers to a sudden change of the current or voltage signal in a short time, which is usually caused by a real physical process, such as switch operation, short-circuit fault, arc discharge, load mutation, etc.

[0072] If the event is determined as a non-false event, it is determined that a transient event exists.

[0073] The transient event can be a precursor of an arc fault or a sudden change of the current, and through analysis of the event, the system can determine whether an arc fault occurs.

[0074] S250, if the transient event exists, determining a target fault detection model according to the at least one time-domain feature and the change interval.

[0075] The transient feature is extracted from the at least one time-domain feature, the load type is determined according to the transient feature and the change interval, and the target fault detection model is determined according to the load type.

[0076] The load type includes a resistive load, an inductive load, a switchable load, a capacitive load, a new energy load, etc.

[0077] The current waveform of each load type has a unique transient feature. The feature of the resistive load in the transient event is a current fluctuation in a short time; the feature of the inductive load in the transient event is a large transient fluctuation of the current waveform at start-up, with a long duration; the feature of the switchable load in the transient event is a current pulse in a short time; the feature of the capacitive load in the transient event is a large current pulse in a short time; and the feature of the new energy load in the transient event is a small high-frequency fluctuation of the current at the initial start-up, continuous and non-periodic change of the current amplitude during operation, and a short-time reverse current pulse when the energy storage system switches from charging to discharging.

[0078] Therefore, the interval duration of the change interval can be determined first, at least one candidate load type is determined according to the interval duration; the time position of the change interval in the time window is determined; and the load type is determined according to the time position and the transient feature in the at least one candidate load type.

[0079] The electrical characteristics and fault modes of different load types are significantly different, so different load types correspond to different fault detection models.

[0080] wherein, the characteristic data of the electric arc is received, the characteristics are analyzed based on the trained fault detection model, and the output result is output, for example, judging that the electric arc is a normal working electric arc, an abnormal electric arc caused by poor contact, a short circuit fault electric arc, etc.

[0081] For example, for a resistive load, the model can adopt a time series network containing a short-time sliding window; for an inductive load, the model can adopt a deep network containing a long-time sequence window; for a switchable load, the model adopts an adaptive network with a mixed time sequence window; for a capacitive load, the model can adopt a convolutional neural network with a high-frequency convolution kernel; for a new energy load, the model can adopt a network combining attention mechanism with Transformer.

[0082] S260, the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features are subjected to feature screening processing and feature fusion processing to obtain target features.

[0083] In one possible embodiment, the feature screening processing and the feature fusion processing of the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features to obtain target features include: determining a first load type corresponding to the current; obtaining monitoring requirements of the first load type; creating a first feature subset, the first feature subset at least including any two different dimension features in the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features; performing feature expansion processing on the first feature subset to obtain a second feature subset, the feature expansion processing being used to expand a single other feature to the current feature subset, the other feature being a feature in the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features, except for the first feature subset; determining a first correlation degree of the second feature subset and the monitoring requirements; determining a feature processing mode of the second feature subset according to a third difference value of the first correlation degree and a preset correlation degree, the feature processing mode including the feature expansion processing and / or feature elimination processing, the feature elimination processing being used to eliminate a single feature in the current feature subset; performing feature processing on the second feature subset according to the feature processing mode to obtain a third feature subset; determining a second correlation degree of the third feature subset and the monitoring requirements; detecting that the number of features in the third feature subset is a preset number, the second correlation degree is greater than or equal to the preset correlation degree, and each of the other features is processed by the feature expansion processing and / or the feature elimination processing; and performing fusion processing on each feature in the third feature subset to obtain the target features.

[0084] wherein, the current load type is determined according to the transient characteristic and the change interval.

[0085] Among them, the monitoring requirements of different load types are different, for example, the monitoring requirements of resistive load include identifying poor contact faults, warning overload risks, detecting short circuit hazards, etc.; the monitoring requirements of capacitive load include detecting partial discharge faults, warning overvoltage breakdown risks, etc.

[0086] Among them, the first feature subset includes at least two-dimensional features, such as two time domain features and one second statistical feature.

[0087] Among them, when calculating the correlation degree of the feature subset and the monitoring requirement by a statistical method such as Pearson correlation coefficient, the monitoring requirement is converted into a feature evaluation index, the evaluation value of the feature subset is determined according to the evaluation index, and the contribution degree of the selected feature subset to meet the monitoring requirement, i.e. the correlation degree, is measured by the evaluation value.

[0088] For example, the feature evaluation index corresponding to the inductive load can include an index for evaluating the feature's ability to capture the starting peak and an index for evaluating the feature's ability to represent the harmonic. For example, the feature evaluation index corresponding to the capacitive load can include an index for evaluating the feature's tracking accuracy for the impact process and an index for evaluating the feature's ability to represent the stable operation.

[0089] Among them, the feature expansion processing is used to select one feature from other features to join the current feature subset, perform feature evaluation on the feature subset after joining according to the feature evaluation index, obtain the evaluation value, and determine the evaluation value as the correlation degree of the feature subset and the detection requirement, so that the correlation degree of the new subset after joining the feature and the monitoring requirement is optimal.

[0090] Among them, the feature expansion processing is used to select two features from other features to join the current feature subset, perform feature evaluation on the feature subset after joining according to the feature evaluation index, obtain the evaluation value, and determine the evaluation value as the correlation degree of the feature subset and the detection requirement, so that the correlation degree of the new subset after joining the feature and the monitoring requirement is optimal.

[0091] Specifically, according to the feature evaluation index, the importance score of each feature in the feature subset is calculated and summed to obtain the evaluation value.

[0092] In a possible embodiment, the feature processing method of the second feature subset is determined based on the third difference between the first correlation and the preset correlation, including: if the third difference is within a first preset range, the feature processing method is the feature expansion processing; if the third difference is within a second preset range, the feature processing method is the feature elimination processing, and the upper limit value of the first preset range is less than the lower limit value of the second preset range; if the third difference is within a third preset range, the feature processing method is the feature elimination processing and the feature expansion processing, and the upper limit value of the second preset range is less than the lower limit value of the third preset range.

[0093] Among them, the feature elimination processing is used to eliminate a feature from the current feature subset, and the feature evaluation of the eliminated feature subset is performed according to the feature evaluation index to obtain an evaluation value, and the evaluation value is determined as the correlation between the feature subset and the detection requirement, so that the correlation between the new subset after the feature is eliminated and the monitoring requirement is optimal.

[0094] Among them, if the above-mentioned feature expansion processing selects two features to be added to the current feature subset, the feature elimination processing can be used to eliminate one or two features from the current feature subset, and the feature evaluation of the eliminated feature subset is performed according to the feature evaluation index to obtain an evaluation value, and the evaluation value is determined as the correlation between the feature subset and the detection requirement, so that the correlation between the new subset after the feature elimination and the monitoring requirement is optimal.

[0095] The maximum value of the first preset range is smaller than the minimum value of the second preset range; the maximum value of the second preset range is smaller than the minimum value of the third preset range, and the three have no overlap and are sorted in ascending order.

[0096] When the third difference is within a first preset range, indicating that the relevance of the feature subset to the monitoring requirement is close to the preset relevance, feature expansion processing is performed to add new relevant feature dimensions to compensate for the shortcomings of the current features and optimize the current relevance. It is allowed to retroactively remove old features after adding new features.

[0097] When the third difference is within the second preset range, it indicates that the relevance of the current feature subset to the monitoring requirements is somewhat below the preset threshold, indicating feature redundancy. Feature elimination is performed to remove weakly relevant or irrelevant features and retain the core features to optimize the current relevance. After eliminating features, better features can be added back.

[0098] Among them, when the third difference is within the third preset range, it indicates that the correlation between the current feature subset and the monitoring requirement is far lower than the preset threshold, and feature expansion processing and feature elimination processing can be performed simultaneously to optimize the current correlation.

[0099] Among them, after obtaining the third feature subset, the second correlation between the third feature subset and the monitoring requirements is determined. If the number of features in the third feature subset is a preset number, and the second correlation is greater than or the preset correlation, and each other feature has been processed by feature expansion and / or feature elimination, then each feature in the third feature subset is fused to obtain the target feature.

[0100] Among them, if the above conditions are not met at the same time, the difference between the second correlation and the preset correlation is calculated, and based on the difference, the feature expansion processing and / or feature elimination processing is repeatedly performed according to the method described above, until the correlation between the feature subset and the monitoring requirements reaches the preset correlation, and the correlation cannot be improved by adding or eliminating a single feature, and the number of features in the feature subset reaches the preset number, then the obtained feature subset is subjected to feature fusion processing to obtain the target feature.

[0101] Among them, feature fusion processing can be determined by weighted summation or by normalization and splicing. For example, features of different magnitudes are normalized to the same interval and then spliced ​​into vectors, retaining their respective information while eliminating magnitude differences to obtain target features.

[0102] It can be seen that in the embodiment of the present application, starting from the load monitoring requirements, an effective feature subset is constructed through feature engineering and its correlation with the monitoring requirements is quantified, providing highly targeted input features for subsequent model analysis.

[0103] In a possible embodiment, the method further includes: if the transient event does not exist, obtaining a second load type triggered by a preceding transient event; determining a target fault detection model based on the second load type; performing feature screening and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the multiple second statistical features to obtain a target feature; and determining a fault detection result based on the target feature and the target fault detection model.

[0104] Among them, if no transient event is detected, the load type triggered and identified by the previous transient event is determined as the current load type, and then the fault detection model is determined, and feature screening and feature fusion processing are performed on the multi-dimensional features. Finally, the obtained target features are input into the target fault detection model, and the fault detection results are output.

[0105] S270: Determine a fault detection result according to the target feature and the target fault detection model.

[0106] The fault detection results may include fault type, fault occurrence time, fault characteristic parameters, fault level, associated load information, etc.

[0107] If an arc fault is detected, the system outputs an alarm message and performs corresponding isolation or control measures based on the load type and detection results, such as isolating the fault area, disconnecting the power supply, etc.

[0108] Different response measures are determined according to the load type and fault detection results, for example, for a resistive load fault, the power supply can be directly cut off; while in the case of an inductive load fault, a more refined control strategy may be needed.

[0109] If no arc fault is detected, the system will continue to monitor the power grid state in real time and update the load classification information, ensuring that the system is always in the best detection state.

[0110] As can be seen, in the embodiments of the present application, the multi-dimensional data is used to determine the transient event, which helps to eliminate false events and improve the accuracy of determination. Then, the matching of the fault detection model is triggered according to the transient event. When performing feature processing, the time domain, frequency domain features and statistical features are screened and combined to eliminate redundant features. Then, the fault detection is performed by combining the target features obtained after feature processing and the target fault detection model obtained by matching, which can make the fault detection result more in line with the actual situation of the load, thereby improving the pertinence and accuracy of fault detection, avoiding the misjudgment or omission problem caused by the mismatch of load characteristics, and improving the efficiency and accuracy of system fault detection.

[0111] In one possible embodiment, please refer to Figure 4 , Figure 4 is a flowchart of a second arc fault detection method provided by the embodiments of the present application, as Figure 4 shown, the current signal in the power distribution network is monitored in real time, and transient event detection is performed based on the change of the current signal. When an abnormal current waveform is detected, load classification is triggered.

[0112] Among them, the load classification uses an event-driven method to identify the load type according to the transient event detection results, and divides the load into resistive load, inductive load and switchable load, ensuring that the appropriate model algorithm is used for arc fault detection.

[0113] Among them, for different load types, arc fault features are extracted through feature extraction and feature selection, and the most discriminative features are selected for arc fault detection to obtain the arc fault result.

[0114] Among them, if an arc fault is detected, the system outputs an alarm message and performs corresponding isolation or control measures based on the load type and detection results; if no fault is detected, the system continues to monitor and perform the next round of data acquisition and analysis.

[0115] In one possible embodiment, please refer to Figure 5, Figure 5 is a flowchart of a third arc fault detection method provided by an embodiment of the present application, as shown in Figure 5 The system captures the current signal in real time in the collection window, and performs transient event monitoring according to the current signal. Specifically, the current signal is analyzed in segments using a sliding window analysis method to detect mutations or abnormal changes in the current signal. When abnormal changes are detected, false events are detected, i.e., false events are deleted according to the duration of abnormal changes to ensure that false positives are eliminated, and the current signal continues to be captured for the next round of analysis.

[0116] If it is determined that a transient event occurs, the load type is identified, and the load type is determined according to the duration of the transient event and the transient characteristics. For example, if the duration is within a first time range and the transient characteristics represent a large transient fluctuation in the current waveform at startup, it is determined that the current load is an inductive load. The arc fault detector is matched according to the inductive load, and load characteristics are extracted and selected to obtain the most discriminative features, which are input into the matched arc fault detector for arc fault detection, and the arc fault detection result is output.

[0117] If no abnormal changes are detected, the pre-load type identified by the previous transient event trigger is determined as the current load type, and the arc fault detector is matched, load characteristics are extracted and selected to obtain the most discriminative features, which are input into the matched arc fault detector for arc fault detection, and the arc fault detection result is output.

[0118] For the above embodiment, please refer to Figure 6 , Figure 6 is a functional unit composition block diagram of an arc fault detection device provided by an embodiment of the present application, as shown in Figure 6As shown, the arc fault detection apparatus 60 comprises: a first determination unit 61 configured to determine a first statistical feature of a current signal in a time window, the current signal comprising an even number of sampling points, the first statistical feature being configured to represent a local offset level and an overall dispersion degree of the current signal; an analysis unit 62 configured to perform waveform mutation analysis on the current signal to obtain a variation interval with the largest fluctuation amplitude in the current signal; an extraction unit 63 configured to extract at least one time domain feature, at least one frequency domain feature, and a plurality of second statistical features of the current signal, the plurality of second statistical features being configured to represent fluctuation amplitudes and morphological deviations of the current signal; a second determination unit 64 configured to determine whether a transient event exists according to the first statistical feature, real-time current data, and the variation interval; a matching unit 65 configured to determine a target fault detection model according to the at least one time domain feature and the variation interval if the transient event exists; a feature processing unit 66 configured to perform feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; and a third determination unit 67 configured to determine a fault detection result according to the target feature and the target fault detection model.

[0119] In one possible embodiment, in terms of determining the first statistical feature of the current signal in the time window, the first determination unit 61 is specifically configured to: determine a variance of the current signal in the time window; split the time window to obtain a first sub-time window and a second sub-time window, a first number of sampling points being the same as a second number of sampling points, the first number of sampling points being a number of sampling points of the current signal in the first sub-time window, the second number of sampling points being a number of sampling points of the current signal in the second sub-time window; determine a first mean value of the current signal in the first sub-time window; and determine a second mean value of the current signal in the second sub-time window; and obtain the first statistical feature according to the first mean value, the second mean value, and the variance.

[0120] In one possible embodiment, in terms of determining whether a transient event exists according to the first statistical feature, real-time current data, and the variation interval, the second determination unit 64 is specifically further configured to: determine a variation amount of the current signal according to the first mean value, the second mean value, the variance, and the real-time current data; determine an interval length of the variation interval if the variation amount is greater than a preset variation amount; and determine whether a transient event exists according to the interval length.

[0121] In one possible embodiment, in determining the amount of change in the current signal based on the first mean, the second mean, the variance and the real-time current data, the second determination unit 64 is specifically used to: determine a first difference between the first mean and the second mean, wherein the first difference is a positive number; determine a ratio of the first difference and the variance; perform a mean operation on the first mean and the second mean to obtain a third mean; determine a second difference between the real-time current data and the third mean, wherein the second difference is a positive number; and determine the product of the second difference and the ratio to obtain the amount of change.

[0122] In one possible embodiment, in performing feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature and the multiple second statistical features to obtain the target feature, the feature processing unit 66 is specifically further used to: determine the first load type corresponding to the current signal; obtain the monitoring requirements of the first load type; create a first feature subset, the first feature subset includes at least the at least one time domain feature, the at least one frequency domain feature and any two different dimensional features of the multiple second statistical features; perform feature expansion processing on the first feature subset to obtain a second feature subset, the feature expansion processing is used to expand a single other feature into the current feature subset, the other feature is the at least one time domain feature, the at least one frequency domain feature and the multiple second statistical features, except for the first feature subset. features; determining a first correlation between the second feature subset and the monitoring requirement; determining a feature processing method for the second feature subset based on a third difference between the first correlation and the preset correlation, the feature processing method including the feature expansion processing and / or feature elimination processing, the feature elimination processing being used to eliminate a single feature in the current feature subset; performing feature processing on the second feature subset according to the feature processing method to obtain a third feature subset; determining a second correlation between the third feature subset and the monitoring requirement; detecting that the number of features in the third feature subset is a preset number, and the second correlation is greater than or equal to the preset correlation, and each of the other features has been processed by the feature expansion processing and / or the feature elimination processing, then performing fusion processing on each feature in the third feature subset to obtain the target feature.

[0123] In a possible implementation, in determining the feature processing mode of the second feature subset according to the third difference value of the first correlation degree and a preset correlation degree, the feature processing unit 66 is specifically configured to: if the third difference value is within a first preset range, the feature processing mode is the feature expansion processing; if the third difference value is within a second preset range, the feature processing mode is the feature elimination processing, an upper limit value of the first preset range is less than a lower limit value of the second preset range; if the third difference value is within a third preset range, the feature processing mode is the feature elimination processing and the feature expansion processing, an upper limit value of the second preset range is less than a lower limit value of the third preset range.

[0124] In a possible implementation, the arc fault detection apparatus 60 is specifically configured to: if the transient event does not exist, acquire a second load type of pre-transient event trigger identification; determine a target fault detection model according to the second load type; perform feature screening processing and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; and determine a fault detection result according to the target feature and the target fault detection model.

[0125] It can be understood that, since the method embodiments and the apparatus embodiments are different presentation forms of the same technical concept, the content of the method embodiments part in the present application should be synchronously adapted to the apparatus embodiments part, which will not be repeated here.

[0126] In the case of an integrated unit, please refer to Figure 7 , Figure 7 is another functional unit composition block diagram of the arc fault detection apparatus provided by the embodiments of the present application, as Figure 7 shown, the arc fault detection apparatus 60 includes: a processing module 602 and a communication module 601. The processing module 602 is configured to control and manage the actions of the arc fault detection apparatus 60, for example, to perform the steps of the first determination unit 61, the analysis unit 62, the extraction unit 63, the second determination unit 64, the matching unit 65, the feature processing unit 66, and the third determination unit 67, and / or to perform other processes of the techniques described herein. The communication module 601 is configured to interact between the arc fault detection apparatus 60 and other devices.

[0127] As shown in Figure 7 , the arc fault detection apparatus 60 can further include a storage module 603, and the storage module 603 is configured to store program codes and data of the arc fault detection apparatus 60.

[0128] The processing module 602 can be a processor or a controller, for example, a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof. The processing module 602 can implement or execute the various exemplary logical blocks, modules and circuits described in connection with the disclosure. The processing module 602 can also be a combination of computing functions, for example, a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0129] The communication module 601 can be a transceiver, an RF circuit or a communication interface. The storage module 603 can be a memory.

[0130] The above method embodiments involve all related content of each scenario, which can be cited to the function description of the corresponding function module, and will not be repeated here. The above arc fault detection device 60 can execute the above arc fault detection method. Figure 2 The arc fault detection method shown in the above embodiment.

[0131] Please refer to Figure 8 , Figure 8 is a structural schematic diagram of an electronic device according to an embodiment of the present application, as Figure 8 shown, the electronic device 800 includes a processor 810, a memory 820, a communication interface 830 and one or more programs 821, the above one or more programs 821 are stored in the above memory, and are configured to be executed by the above processor, the program includes part or all steps of any arc fault detection method described in the above method embodiments when executed, the processor, the memory and the communication interface are connected to each other and complete the communication work between each other.

[0132] The memory can be a volatile memory such as a dynamic random access memory (DRAM), or a non-volatile memory such as a mechanical hard disk. The above memory is used to store a set of executable program codes, and the above processor is used to call the executable program codes stored in the memory, which can execute part or all steps of any arc fault detection method described in the above arc fault detection method embodiments.

[0133] It can be seen that the electronic device 800 described in the embodiments of the present application first determines the first statistical feature of the current signal in the time window, the current signal includes an even number of sampling points, and the first statistical feature is used to represent the local offset level and the overall dispersion degree of the current signal. Then, the waveform mutation analysis is performed on the current signal to obtain the change interval with the largest fluctuation amplitude in the current signal. Then, at least one time domain feature, at least one frequency domain feature and a plurality of second statistical features of the current signal are extracted, and the plurality of second statistical features are used to represent the fluctuation amplitude and the morphological deviation of the current signal. Then, according to the first statistical feature, the real-time current data and the change interval, it is determined whether there is a transient event. If the transient event exists, the target fault detection model is determined according to the at least one time domain feature and the change interval. Then, the at least one time domain feature, the at least one frequency domain feature and the plurality of second statistical features are subjected to feature screening processing and feature fusion processing to obtain a target feature. Finally, according to the target feature and the target fault detection model, a fault detection result is determined.

[0134] The present application determines the transient event by multi-dimensional data, which is beneficial to eliminate false events and improve the accuracy of determination. Then, the matching of the fault detection model is triggered according to the transient event. When performing feature processing, the time domain, frequency domain features and statistical features are screened and combined to eliminate redundant features. Then, the target feature obtained after the feature processing and the target fault detection model obtained by matching are combined for fault detection, which can make the fault detection result more in line with the actual situation of the load, thereby improving the pertinence and accuracy of fault detection, avoiding the misjudgment or omission problem caused by the mismatch of load characteristics, and improving the efficiency and accuracy of power system fault detection.

[0135] The embodiments of the present application also provide a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program causes a computer to execute part or all of the steps of any method described in the above method embodiments. The above computer includes an electronic device.

[0136] The embodiments of the present application also provide a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute part or all of the steps of any method described in the above method embodiments. The computer program product can be a software installation package, and the computer includes an electronic device.

[0137] It should be noted that, for the methods described above, each of the individual steps is described in a certain order, but the skilled person should understand that the present application is not limited by the order of the steps, and some steps can be performed in other orders or simultaneously. In addition, the skilled person should understand that the embodiments described in the specification are optional embodiments, and the steps and modules involved are not necessarily essential to the present application.

[0138] In the above embodiments, the description of each embodiment focuses on different aspects, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0139] In the several embodiments provided by the present application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely schematic, and the division of the units is merely a logical function division. In actual implementation, another division manner can be adopted, for example, a plurality of units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical or other forms.

[0140] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

[0141] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software program module.

[0142] If the integrated unit is realized in the form of a software program module and sold or used as an independent product, it can be stored in a computer readable memory. Based on this understanding, the technical solutions of the present application essentially or related parts of the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0143] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by a program instructing relevant hardware, and the program can be stored in a computer readable memory, which can include a flash disk, a read-only memory, a random access memory, a magnetic disk or an optical disk, etc.

[0144] The embodiments of the present application are described in detail above, and specific examples are applied herein to describe the principles and embodiments of the present application. The above description of the embodiments is only used to help understand the method of the present application and its core idea; meanwhile, for those skilled in the art, according to the idea of the present application, the specific embodiments and application scope will be changed, and the above description of the embodiments should not be understood as a limitation of the present application.

Claims

1. A method for detecting an arc fault, characterized in that: include: determining a first statistical feature of a current signal within a time window, the current signal including an even number of sampling points, the first statistical feature being used to characterize a local offset level and an overall discreteness of the current signal; Performing waveform mutation analysis on the current signal to obtain a variation interval with the largest fluctuation amplitude in the current signal; extracting at least one time domain feature, at least one frequency domain feature, and a plurality of second statistical features of the current signal, wherein the plurality of second statistical features are used to characterize the fluctuation amplitude and morphological deviation of the current signal; determining whether a transient event exists based on the first statistical feature, the real-time current data, and the variation interval; If the transient event exists, determining a target fault detection model according to the at least one time domain feature and the variation interval; performing feature screening and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; A fault detection result is determined according to the target feature and the target fault detection model.

2. The method according to claim 1, characterized in that The determining of the first statistical feature of the current signal within the time window includes: determining a variance of the current signal within the time window; Splitting the time window to obtain a first sub-time window and a second sub-time window, wherein the number of first sampling points is the same as the number of second sampling points, the first number of sampling points being the number of sampling points of the current signal in the first sub-time window, and the second number of sampling points being the number of sampling points of the current signal in the second sub-time window; determining a first mean value of the current signal within the first sub-time window; and determining a second mean value of the current signal within the second sub-time window; The first statistical feature is obtained according to the first mean, the second mean, and the variance.

3. The method according to claim 2, characterized in that The determining whether a transient event exists according to the first statistical feature, the real-time current data, and the change interval includes: determining a change in the current signal according to the first mean, the second mean, the variance, and the real-time current data; If it is detected that the change amount is greater than a preset change amount, then determining the interval length of the change interval; Determine whether a transient event exists based on the interval duration.

4. The method according to claim 3, characterized in that The determining the variation of the current signal according to the first mean, the second mean, the variance, and the real-time current data includes: determining a first difference between the first mean and the second mean, wherein the first difference is a positive number; determining a ratio of the first difference to the variance; performing a mean operation on the first mean and the second mean to obtain a third mean; determining a second difference between the real-time current data and the third mean, wherein the second difference is a positive number; The product of the second difference and the proportional value is determined to obtain the change.

5. The method according to claim 1, wherein The performing feature screening and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature includes: determining a first load type corresponding to the current signal; Obtaining monitoring requirements for the first load type; Creating a first feature subset, where the first feature subset includes at least the at least one time domain feature, the at least one frequency domain feature, and any two different dimensional features among the plurality of second statistical features; performing feature expansion processing on the first feature subset to obtain a second feature subset, wherein the feature expansion processing is used to expand a single other feature into the current feature subset, the other feature being a feature other than the first feature subset among the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features; determining a first correlation between the second feature subset and the monitoring requirement; determining, based on a third difference between the first correlation and a preset correlation, a feature processing method for the second feature subset, the feature processing method including the feature expansion processing and / or feature elimination processing, the feature elimination processing being used to eliminate a single feature in the current feature subset; Performing feature processing on the second feature subset according to the feature processing method to obtain a third feature subset; determining a second correlation between the third feature subset and the monitoring requirement; If it is detected that the number of features in the third feature subset is a preset number, and the second correlation is greater than or equal to the preset correlation, and each of the other features has been processed by the feature expansion and / or the feature elimination, each feature in the third feature subset is fused to obtain the target feature.

6. The method according to claim 5, characterized in that The determining, based on a third difference between the first correlation and a preset correlation, a feature processing method for the second feature subset includes: If the third difference is within a first preset range, the feature processing method is the feature expansion processing; If the third difference is within a second preset range, the feature processing method is the feature elimination processing, and the upper limit of the first preset range is less than the lower limit of the second preset range; If the third difference is within a third preset range, the feature processing method is the feature elimination processing and the feature expansion processing, and the upper limit value of the second preset range is less than the lower limit value of the third preset range.

7. The method according to any one of claims 1 to 6, characterized in that The method further comprises: If the transient event does not exist, obtaining a second load type identified by a preceding transient event trigger; determining a target fault detection model according to the second load type; performing feature screening and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; A fault detection result is determined according to the target feature and the target fault detection model.

8. An arc fault detection device, characterized in that: include: a first determining unit, configured to determine a first statistical feature of a current signal within a time window, the current signal including an even number of sampling points, the first statistical feature being used to characterize a local offset level and an overall discreteness of the current signal; An analysis unit, configured to perform waveform mutation analysis on the current signal to obtain a variation interval with the largest fluctuation amplitude in the current signal; an extraction unit, configured to extract at least one time domain feature, at least one frequency domain feature, and a plurality of second statistical features of the current signal, wherein the plurality of second statistical features are used to characterize the fluctuation amplitude and morphological deviation of the current signal; a second determining unit, configured to determine whether a transient event exists based on the first statistical feature, the real-time current data, and the change interval; a matching unit, configured to determine a target fault detection model based on the at least one time domain feature and the variation interval if the transient event exists; a feature processing unit, configured to perform feature screening and feature fusion processing on the at least one time domain feature, the at least one frequency domain feature, and the plurality of second statistical features to obtain a target feature; The third determining unit is configured to determine a fault detection result according to the target feature and the target fault detection model.

9. An electronic device, characterized in that: The device comprises: A memory, a processor, and an executable program code stored in the memory and executable on the processor, wherein the processor executes the steps of the arc fault detection method according to any one of claims 1 to 7 when executing the executable program code.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores executable program code, wherein the executable program code includes execution instructions, and the execution instructions are used to execute the steps of the arc fault detection method according to any one of claims 1 to 7.

Images