Low-voltage fault arc identification method and apparatus, medium and device

Abstract

Embodiments of the present disclosure provide a low-voltage fault arc identification method and apparatus, a medium and a device. The method comprises: using a period difference method to perform feature extraction on an acquired low-frequency signal within 3kHz of an electricity meter, and determining a low-frequency fault arc time domain feature vector; using generalized S-transform to perform feature extraction on an acquired intermediate-frequency current signal within 100kHz of the electricity meter, and determining an intermediate-frequency fault arc energy feature; inputting the low-frequency fault arc time domain feature vector and the intermediate-frequency fault arc energy feature into a pre-trained neural network identification model, and outputting fault arc signal presence information; and on the basis of the fault arc signal presence information, determining whether a fault arc signal is present or not.

Description

A low-voltage fault arc identification method, device, medium and equipment

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This disclosure is based on the Chinese patent application with application number 202311602774.7, application date November 28, 2023, and application name “A method, device, medium and equipment for identifying low-voltage fault arc”, and claims the priority of the Chinese patent application. The entire content of the Chinese patent application is hereby introduced into this disclosure as a reference. Technical Field

[0003] The present disclosure relates to, but is not limited to, the technical field of low-voltage arc fault protection, and more specifically, to a low-voltage arc fault identification method, apparatus, medium, and equipment. Background Art

[0004] A safe and reliable power supply is fundamental to national political security, social stability, and economic development. According to government statistics, approximately 470,000 electrical fires occurred in my country between 2017 and 2021, accounting for over 30% of the nation's total fire deaths and injuries. Nearly 70% of these fires were caused by arc faults, posing a serious threat to public life and property and jeopardizing the safe and stable operation of the power grid. The State Council's Work Safety Committee has issued the "Notice on Comprehensively Managing Electrical Fires" and the "Notice on Specialized Rectification of Major Fire Risks in High-Rise Buildings," requiring "all regions to strengthen the application of technologies such as intelligent fire monitoring and early warning, comprehensive early warning and prevention of electrical fires, and cloud service platforms for fire safety management." Currently, existing technologies, based on GB / T31143-2014 "General Requirements for Arc Fault Protection Devices (AFDD)" and GB14287.4-2014 "Electrical Fire Monitoring Systems - Part 4: Arc Fault Detectors," have developed arc fault circuit breakers and arc fault detectors, respectively. These are primarily used in low-voltage 220V circuits to monitor and disconnect the circuit when an arc fault occurs. However, due to their relatively simple technical requirements and limited protection range, they are prone to high rates of missed alarms and false alarms, resulting in limited widespread adoption. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the embodiments of the present disclosure provide a low-voltage arc fault identification method, device, medium and equipment.

[0006] According to one aspect of an embodiment of the present disclosure, a low-voltage arc fault identification method is provided, comprising:

[0007] The period difference method is used to extract the features of the low-frequency signal within the 3kHz range of the collected electric energy meter to determine the time domain feature vector of the low-frequency fault arc.

[0008] The generalized S transform is used to extract the features of the 100kHz intermediate frequency current signal collected from the electric energy meter to determine the energy characteristics of the intermediate frequency fault arc.

[0009] The low-frequency fault arc time domain feature vector and the medium-frequency fault arc energy feature are input into the pre-trained neural network recognition model to output the existence information of the fault arc signal;

[0010] Determine whether an arc fault signal exists based on the arc fault signal existence information.

[0011] In some embodiments, a period difference method is used to extract features from the collected low-frequency signal within the 3 kHz range of the electric energy meter to determine the time domain feature vector of the low-frequency fault arc, including:

[0012] Collect low-frequency signals within the 3kHz range of the energy meter, including low-frequency current, voltage, and event information;

[0013] Calculate the low-frequency power consumption characteristics of any period of the low-frequency signal to form a low-frequency time domain feature vector of the low-frequency signal;

[0014] The difference between adjacent period feature vectors is compared based on the low-frequency time domain feature vector to form the low-frequency fault arc time domain feature vector.

[0015] In some embodiments, comparing the differences of adjacent period feature vectors based on the low-frequency time domain feature vector to form the low-frequency fault arc time domain feature vector includes:

[0016] Determining signal similarity of multiple periodic signals based on calculated current differences between adjacent periods;

[0017] performing normalization processing on the signal similarity to determine normalized signal similarity;

[0018] Calculate the period difference based on the signal similarity and the normalized signal similarity;

[0019] According to the period difference, the time domain characteristic vector of the low-frequency fault arc is determined.

[0020] In some embodiments, a generalized S-transform is used to extract features from a 100kHz intermediate frequency current signal collected from an electric energy meter to determine the energy features of the intermediate frequency fault arc, including:

[0021] Collect the 100kHz intermediate frequency current signal of the electric energy meter and perform noise reduction and filtering;

[0022] Perform harmonic analysis on the intermediate frequency current signal to determine the harmonic content;

[0023] For intermediate frequency current signals with harmonic content greater than a preset percentage, a hyperbolic Gaussian window function is used to perform harmonic elimination processing;

[0024] For medium-frequency current signals with harmonic content not exceeding a preset percentage, a Gaussian window function is used for harmonic elimination processing;

[0025] The generalized S transform is used to extract the current characteristics of the intermediate frequency current signal after harmonic elimination and determine the energy characteristics of the intermediate frequency fault arc.

[0026] In some embodiments, the hyperbolic Gaussian window function is:

[0027] When t≥τ,

[0028] When t<τ,

[0029] Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and They represent the taper of the second half of the Gaussian curve and the taper of the first half of the Gaussian curve, respectively, and f represents the frequency.

[0030] In some embodiments, the training process of the neural network recognition model is:

[0031] Collecting current signal sample data, wherein the current signal sample data includes current signal sample data of multiple low-voltage loads in normal and fault states and current signal sample data of a state in which at least two low-voltage loads are turned on and working simultaneously;

[0032] Expanding the current signal sample data by using a cyclic noise addition method to determine the expanded current signal sample data;

[0033] The expanded current signal sample data is randomly divided into a training set and a test set according to a preset ratio;

[0034] Train the neural network recognition model based on the training set and the test set.

[0035] According to another aspect of the embodiments of the present disclosure, a low-voltage arc fault identification device is provided, comprising:

[0036] The first extraction module is configured to extract features from the low-frequency signal within the 3kHz range of the collected electric energy meter using a period difference method to determine a time-domain feature vector of the low-frequency fault arc;

[0037] The second extraction module is configured to use a generalized S-transform to extract features from the collected 100kHz intermediate frequency current signal of the electric energy meter to determine the energy characteristics of the intermediate frequency fault arc;

[0038] An output module is configured to input the low-frequency fault arc time domain feature vector and the medium-frequency fault arc energy feature into a pre-trained neural network recognition model and output the presence information of the fault arc signal;

[0039] The determination module is configured to determine whether an arc fault signal exists according to the arc fault signal existence information.

[0040] According to another aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, wherein the storage medium stores a computer program, and the computer program is configured to execute the method described in any one of the above aspects of the embodiment of the present disclosure.

[0041] According to another aspect of an embodiment of the present disclosure, an electronic device is provided, comprising: a processor; a memory configured to store instructions executable by the processor; and the processor configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of the above aspects of the embodiment of the present disclosure.

[0042] Therefore, the present invention identifies arc faults when they occur in the customer-side circuit, can perform effective measurement and diagnosis, prevent electrical fire accidents, and provide technical support for protecting electricity safety, maintaining electricity supply and use order, and ensuring safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] A more complete understanding of exemplary embodiments of the present disclosure may be obtained by referring to the following drawings:

[0044] FIG1 is a schematic flow chart of a low-voltage arc fault identification method provided by an embodiment of the present disclosure;

[0045] FIG2 is another schematic flow chart of a low-voltage arc fault identification method according to an embodiment of the present disclosure;

[0046] FIG3 is a schematic diagram of the basic architecture of a two-dimensional convolutional neural network driven by filtered fault arc data provided by an embodiment of the present disclosure;

[0047] FIG4 is a schematic diagram of a low-voltage arc fault broadband measurement and diagnosis device for an electric energy metering device provided by an embodiment of the present disclosure;

[0048] FIG5 is a schematic structural diagram of a low-voltage arc fault identification device provided by an embodiment of the present disclosure;

[0049] FIG6 shows the structure of an electronic device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] Below, the exemplary embodiments according to the present disclosure will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure, and it should be understood that the present disclosure is not limited to the exemplary embodiments described herein.

[0051] It should be noted that the relative arrangement of components and steps, the numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure unless specifically stated otherwise.

[0052] Those skilled in the art will understand that the terms "first" and "second" in the embodiments of the present disclosure are only used to distinguish different steps, devices or modules, and do not represent any specific technical meanings, nor do they indicate a necessary logical order between them.

[0053] It should also be understood that in the embodiments of the present disclosure, “a plurality of” may refer to two or more than two, and “at least one” may refer to one, two, or more than two.

[0054] It should also be understood that any component, data or structure mentioned in the embodiments of the present disclosure can generally be understood as one or more, unless explicitly limited or otherwise indicated in the context.

[0055] In addition, the term "and / or" in this disclosure is merely a description of the association relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this disclosure generally indicates that the related objects are in an "or" relationship.

[0056] It should also be understood that the description of the various embodiments in this disclosure focuses on the differences between the various embodiments, and the same or similar aspects thereof can be referenced with each other. For the sake of brevity, they will not be described one by one.

[0057] At the same time, it should be understood that for the convenience of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship.

[0058] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0059] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0060] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0061] The embodiments of the present disclosure can be applied to electronic devices such as terminal devices, computer systems, and servers, and can operate in conjunction with numerous other general-purpose or specialized computing system environments or configurations. Examples of well-known terminal devices, computing systems, environments, and / or configurations suitable for use with terminal devices, computer systems, servers, and other electronic devices include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network personal computers, minicomputer systems, mainframe computer systems, and distributed cloud computing technology environments including any of the above systems, among others.

[0062] Electronic devices such as terminal devices, computer systems, and servers can be described in the general context of computer system-executable instructions (such as program modules) executed by a computer system. Generally, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in a distributed cloud computing environment, where tasks are performed by remote processing devices linked via a communication network. In a distributed cloud computing environment, program modules can be located on local or remote computing system storage media, including storage devices.

[0063] Exemplary Methods

[0064] FIG1 is a flow chart of a low-voltage arc fault identification method provided by an embodiment of the present disclosure. The present disclosure embodiment can be applied to electronic devices. As shown in FIG1 , the low-voltage arc fault identification method 100 includes the following steps:

[0065] Step 101 : extracting features from the collected low-frequency signal within the 3 kHz range of the electric energy meter using a period difference method to determine a time-domain feature vector of the low-frequency fault arc.

[0066] In some embodiments, a period difference method is used to extract features from the collected low-frequency signal within the 3 kHz range of the electric energy meter to determine the time domain feature vector of the low-frequency fault arc, including:

[0067] Collect low-frequency signals within the 3kHz range of the energy meter, including low-frequency current, voltage, and event information;

[0068] Calculate the low-frequency power consumption characteristics of any period of the low-frequency signal to form a low-frequency time domain feature vector of the low-frequency signal;

[0069] The difference between adjacent period feature vectors is compared based on the low-frequency time domain feature vector to form the low-frequency fault arc time domain feature vector.

[0070] In some embodiments, comparing the differences of adjacent period feature vectors based on the low-frequency time domain feature vector to form the low-frequency fault arc time domain feature vector includes:

[0071] Determining signal similarity of multiple periodic signals based on calculated current differences between adjacent periods;

[0072] performing normalization processing on the signal similarity to determine normalized signal similarity;

[0073] Calculate the period difference based on the signal similarity and the normalized signal similarity;

[0074] According to the period difference, the time domain characteristic vector of the low-frequency fault arc is determined.

[0075] In some embodiments, a method for feature extraction and feature selection of low-voltage arc faults for electric energy metering devices is proposed. A broadband measurement and diagnosis method is used to extract and screen broadband arc fault signals. During implementation, referring to FIG2 , step S201 includes collecting electricity meter power consumption information, such as low-frequency current, voltage, and event information. For any period of the signal, low-frequency power consumption characteristics such as harmonics and zero-crossing characteristics, current and voltage peaks, standard deviations, crest factors, and pulse factors are calculated to form a time-domain feature vector of the signal. The feature vectors of adjacent periods are compared to form a time-domain feature vector of the low-frequency arc fault. Step S202 includes collecting an intermediate-frequency current signal, analyzing its time-frequency and energy-domain feature information, extracting features such as high-order harmonic content, eliminating false alarm interference from nonlinear loads in adjacent periods, and strengthening the main frequency band features of the arc fault to form an intermediate-frequency arc fault energy feature. Step S203 includes extracting and training features based on a period difference method, a generalized S-transform, a hyperbolic Gaussian window function, and a two-dimensional convolutional neural network to form a comprehensive diagnostic strategy for low-voltage arc faults.

[0076] In some embodiments, the current and voltage signals of the household circuit are first obtained for the low-frequency sampling frequency within the range of 3kHz of the electric energy meter. For any cycle of the signal, its harmonics and zero-crossing characteristics, current and voltage peaks, standard deviations, peak factors, pulse factors and other low-frequency power consumption characteristics are calculated. The average value of the absolute value of the difference between the waveform characteristics of adjacent cycles is used to compare the difference between the waveforms of adjacent cycles before and after the fault. Taking the current of the electric energy meter as an example:

[0077] Assume that the number of current sampling points in a single cycle is N, the number of cycles is n, and the sampling values ​​are i k-1 ,i k ,i k+1 : △i k1 (j) = i k+1 (j)-i k (j), j=1,2,…,N; k=2,3,…,n (1); △i k2 (j) = i k (j)-i k-1 (j), j=1,2,…,N; k=2,3,…,n (2); △i k3 (j) = i k+1 (j)-i k-1 (j), j=1,2,…,N; k=2,3,…,n (3);

[0078] Using δ k Compare the similarity of the current waveforms of three adjacent cycles and calculate the current difference △i between adjacent cycles k (j) Average of absolute values:

[0079] At the same time, in order to overcome the problem of sampling value deviation, δ k Perform normalization processing.

[0080] The period difference is:

[0081] The randomness of the arc is effectively utilized, and the time domain characteristic vectors of the current before and after the fault are compared through the period difference, thereby reflecting the fault arc characteristics under low-frequency acquisition conditions.

[0082] Step 102: Use generalized S-transformation to extract features from the collected 100kHz intermediate frequency current signal of the electric energy meter to determine the energy features of the intermediate frequency fault arc.

[0083] In some embodiments, a generalized S-transform is used to extract features from a 100kHz intermediate frequency current signal collected from an electric energy meter to determine the energy features of the intermediate frequency fault arc, including:

[0084] Collect the 100kHz intermediate frequency current signal of the electric energy meter and perform noise reduction and filtering;

[0085] Perform harmonic analysis on the intermediate frequency current signal to determine the harmonic content;

[0086] For intermediate frequency current signals with harmonic content greater than a preset percentage, a hyperbolic Gaussian window function is used to perform harmonic elimination processing;

[0087] For medium-frequency current signals with harmonic content not exceeding a preset percentage, a Gaussian window function is used for harmonic elimination processing;

[0088] The generalized S transform is used to extract the current characteristics of the intermediate frequency current signal after harmonic elimination and determine the energy characteristics of the intermediate frequency fault arc.

[0089] In some embodiments, the hyperbolic Gaussian window function is:

[0090] When t≥τ,

[0091] When t<τ,

[0092] Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and They represent the taper of the second half of the Gaussian curve and the taper of the first half of the Gaussian curve, respectively, and f represents the frequency.

[0093] For example, as shown in Figure 2, the line current signal within the intermediate frequency current of 100kHz is sampled and noise is reduced to the 10kHz frequency band for signal time-frequency domain feature analysis. In order to minimize the leakage and diffusion of spectral energy outside the actual signal component, the generalized S transform with frequency asymmetry of the window function is introduced to perform current feature analysis and extraction.

[0094] where is defined as the parameter that determines the frequency size of the window function. To ensure the reversibility of the S transform, it is necessary to satisfy:

[0095] Due to the high harmonic distortion rate of nonlinear loads, the extraction of overall signal features will be somewhat disturbed. To eliminate this effect, a hyperbolic Gaussian window function is selected to process certain nonlinear loads with a harmonic content greater than 40%, while ensuring that the key characteristic information of nonlinear load faults can still be obtained after subsequent signal noise reduction processing. The hyperbolic Gaussian window function uses two different half-Gaussian windows, which gives the signal different tapers in the front and back directions. It is defined as:

[0096] When t≥τ,

[0097] When t<τ,

[0098] Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and where represents the taper of the second and first half of the Gaussian curve, respectively, and f represents frequency. In actual analysis, the Gaussian window is narrowed by decreasing the value. Decreasing the value improves the temporal resolution of the signal's onset, while increasing it improves the frequency domain resolution of periodic signals outside the normal monitoring range.

[0099] A hyperbolic Gaussian window with a narrower main lobe width and smaller spectral tailing is used to process nonlinear current signals with a harmonic content greater than 40%, which improves the accuracy of intermediate frequency arc signal recognition to a certain extent.

[0100] Step 103: input the low-frequency fault arc time domain feature vector and the medium-frequency fault arc energy feature into a pre-trained neural network recognition model, and output the information of the presence of the fault arc signal.

[0101] In some embodiments, the training process of the neural network recognition model is:

[0102] Collecting current signal sample data, wherein the current signal sample data includes current signal sample data of multiple low-voltage loads in normal and fault states and current signal sample data of a state in which at least two low-voltage loads are turned on and working simultaneously;

[0103] Expanding the current signal sample data by using a cyclic noise addition method to determine the expanded current signal sample data;

[0104] The expanded current signal sample data is randomly divided into a training set and a test set according to a preset ratio;

[0105] Train the neural network recognition model based on the training set and the test set.

[0106] In some embodiments, the training process of the neural network recognition model is:

[0107] A total of 470 samples of 12 loads in normal and faulty states were collected. In addition, 500 current signal samples of at least two loads working simultaneously were randomly turned on, totaling 970 samples. The samples were further expanded by cyclic noise addition. Finally, the samples were randomly divided into training set and test set at a ratio of approximately 20:1.

[0108] In this disclosed embodiment, a two-dimensional convolutional neural network (CNN) was used to train and classify the time-frequency signal features. Analysis of the time-frequency spectrum and harmonic content revealed that the time-frequency features were all within 10 kHz, so the time-frequency signal was subjected to noise reduction and filtering.

[0109] The generalized S-transform provides the data foundation for subsequent signal recognition and analysis, and the signal is further normalized before input into the network. The image pixel size is set to 28×28, and both the input and output are three-dimensional data. The basic architecture of the two-dimensional convolutional neural network driven by filtered fault arc data is shown in Figure 3. It includes an input layer 31, at least one convolutional layer 32, at least one maximum pooling layer 33, a fully connected layer 34, and an output layer 35. The input layer 31 inputs three-dimensional data with 28×28 image pixels. The convolution kernel size in the convolution layer 32 is 3x3, and the maximum pooling layer 33 uses a pooling window size of 2x2 during the pooling operation. Activation functions can increase nonlinearity within the neural network to improve the network's approximation capabilities.

[0110] Among them, Relu is selected as the normalized activation function between the convolution layer 32 and the maximum pooling layer 33, and the batch normalization process ensures that the output of each layer has the same distribution.

[0111] The fully connected layer 34 selects Softmax as the final output of the activation function, which can be calculated as follows:

[0112] where z i is the output value of the i-th node, and C is the number of output nodes, that is, the number of classification categories. The output of the output layer is between 0 and 1, and its value can represent the probability of the classification result being 0 or 1. 0 indicates that there is no fault arc condition on this line, and no alarm is issued; 1 indicates that a fault arc condition has occurred on this line, and an alarm is issued.

[0113] Finally, set the training options for the network, specify the SGDM solver, set the maximum number of iterations to 10, and set the global learning rate to e -3 ,After 10 rounds of 100 iterative training processes, the model accuracy reached more than 95% and the loss rate dropped below 0.1.

[0114] Step 104 : Determine whether there is an arc fault signal according to the arc fault signal existence information.

[0115] In some embodiments, the output of the present disclosure is "whether an arc signal exists", such as 0 for no and 1 for yes.

[0116] In addition, as shown in FIG4 , the present disclosure can implement the above-mentioned method through a hardware device. The hardware device mainly includes an IoT electric energy meter 4 and a fault arc measurement and diagnosis module 5. The IoT electric energy meter 1 includes a voltage sampling unit 41, a current sampling unit 42, an electric energy metering unit 43, and an MCU 44. The fault arc measurement and diagnosis module 5 includes a broadband arc signal acquisition unit 51, an LPF low-pass filter 52, a PGA 53, an ADC 54, and an arc diagnosis algorithm register 55.

[0117] Based on the 0-3kHz current, voltage and other low-frequency signals of the IoT electricity meter, characteristic parameters such as peak value, standard deviation, kurtosis, root mean square, peak factor, pulse factor and margin factor are calculated to identify the time when the fault arc occurs and the fault status.

[0118] At the same time, the 100kHz intermediate frequency arc energy characteristics are acquired by combining the broadband arc signal acquisition unit 51, the LPF low-pass filter 52, the PGA 53 gain amplifier, the ADC 54 sampling module, the DSP arc diagnosis algorithm register 55 and other units.

[0119] Its main links include: based on the voltage acquisition unit 41 and current acquisition unit 42 in the IoT electricity meter, low-frequency signals such as current and voltage can be sampled at 0-3kHz, and at the same time, the wide-band arc signal acquisition unit 51 can be combined to sample medium-frequency current signals at 0-100kHz.

[0120] The electric energy metering unit 43 calculates characteristic parameters such as peak value, standard deviation, kurtosis, root mean square, peak factor, pulse factor and margin factor for the low-frequency sampled current and voltage signals, and outputs the fault arc characteristic value.

[0121] For the intermediate frequency sampling current signal, the signal exceeding the ADC bandwidth is first filtered out through the LPF low-pass filter 52 to obtain the intermediate frequency arc current signal of 0-10kHz; secondly, the gain amplifier 53 is used to determine the gain amplification range of 0-56dB and the gain attenuation range of 0.125-1dB, so as to flexibly control and adjust the weak fault arc signal to the appropriate range; then, the analog current signal is converted into a digital signal through the ADC 54 sampling module, and finally sent to the DSP arc diagnosis algorithm register 55 for lightweight arc time-frequency domain diagnosis operation, and finally outputs the arc characteristic values ​​extracted in the time domain and frequency domain respectively.

[0122] Finally, by using MCU 44 to perform parallel analysis on the low-frequency and medium-frequency combined sampling, the accuracy of arc fault diagnosis is improved through double protection, forming a comprehensive arc fault diagnosis strategy.

[0123] When an arc fault occurs in a customer-side circuit, the disclosed embodiment can perform effective measurement and diagnosis to prevent electrical fire accidents, thereby providing technical support for protecting electricity safety, maintaining the order of electricity supply and use, and ensuring the safe and stable operation of the power grid.

[0124] In summary, compared with the prior art, the embodiments of the present disclosure have the following beneficial effects:

[0125] 1. This method adopts the wide-frequency domain characteristic analysis of the fault arc, and ensures the time domain resolution of the signal through the low-frequency characteristics, thereby ensuring the diagnostic sensitivity and accuracy. It effectively overcomes the diagnostic interference of nonlinear loads on the fault arc signal through the intermediate frequency characteristics, introduces the generalized S transform with frequency asymmetry of the window function to reduce the leakage and diffusion of spectral energy outside the actual signal component, and uses the Gaussian hyperbolic window function to enhance the main frequency band characteristics of the fault arc, effectively eliminating the false alarm interference of non-linear loads.

[0126] 2. The device is based on the low-frequency electric energy sampling circuit of the electric energy meter, which effectively reduces the R&D and application costs of the fault arc measurement and diagnosis device. Combined with the arc medium-frequency measurement and diagnosis unit, it expands the arc measurement and diagnosis channels, improves the arc diagnosis effect, and saves and optimizes computing resources.

[0127] 3. This method has certain universality and portability. It can be combined not only with electricity meters, but also with Bluetooth circuit breakers, user smart sockets, etc.

[0128] 4. The device has broad application prospects and can be used in key low-voltage monitoring areas such as government offices, urban commercial streets, and areas where electrical fires occur frequently. It can effectively improve the safety and reliability of low-voltage AC system operation, promote the efficiency of low-voltage power safety status inspections, and provide strong technical support for protecting power safety, maintaining power supply and use order, ensuring safe and stable operation of the power grid, and saving the country and people from loss of life and property.

[0129] Exemplary devices

[0130] FIG5 is a schematic diagram of the structure of a low-voltage arc fault identification device provided by an embodiment of the present disclosure. As shown in FIG5 , the device 500 includes:

[0131] The first extraction module 510 is configured to extract features from the low-frequency signal within the 3kHz range of the collected electric energy meter using a period difference method to determine a time-domain feature vector of the low-frequency fault arc;

[0132] The second extraction module 520 is configured to use a generalized S-transform to extract features from the collected 100kHz intermediate frequency current signal of the electric energy meter to determine the intermediate frequency fault arc energy feature;

[0133] The output module 530 is configured to input the low-frequency arc fault time domain feature vector and the medium-frequency arc fault energy feature into a pre-trained neural network recognition model, and output the arc fault signal existence information;

[0134] The determination module 540 is configured to determine whether an arc fault signal exists according to the arc fault signal existence information.

[0135] In some embodiments, the first extraction module 510 includes:

[0136] The first acquisition submodule is configured to acquire low-frequency signals within a 3kHz range from the electric energy meter, wherein the low-frequency signals include: low-frequency current, voltage, and event information;

[0137] A construction submodule is configured to calculate the low-frequency power consumption characteristics of any period of the low-frequency signal to form a low-frequency time domain feature vector of the low-frequency signal;

[0138] The forming submodule is configured to compare the differences of adjacent period feature vectors based on the low-frequency time domain feature vector to form a low-frequency fault arc time domain feature vector.

[0139] In some embodiments, a submodule is formed, comprising:

[0140] a first determining unit configured to determine signal similarity of a plurality of periodic signals based on the calculated current differences between adjacent periods;

[0141] a second determining unit configured to perform normalization processing on the signal similarity and determine a normalized signal similarity;

[0142] a calculation unit configured to calculate a period difference based on the signal similarity and the normalized signal similarity;

[0143] The third determining unit is configured to determine a time domain characteristic vector of the low-frequency fault arc according to the period difference.

[0144] In some embodiments, the second extraction module 520 includes:

[0145] The second acquisition submodule is configured to collect the intermediate frequency current signal of the electric energy meter within 100kHz and perform noise reduction and filtering processing;

[0146] a determination submodule, configured to perform harmonic analysis on the intermediate frequency current signal to determine harmonic content;

[0147] The first processing submodule is configured to use a hyperbolic Gaussian window function to perform harmonic elimination processing on the intermediate frequency current signal having a harmonic content greater than a preset percentage;

[0148] The second processing submodule is configured to use a Gaussian window function to perform harmonic elimination processing on the intermediate frequency current signal having a harmonic content not greater than a preset percentage;

[0149] The extraction submodule is configured to use a generalized S-transform to extract current features from the intermediate frequency current signal after harmonic elimination processing, and determine the energy features of the intermediate frequency fault arc.

[0150] In some embodiments, the hyperbolic Gaussian window function is:

[0151] When t≥τ,

[0152] When t<τ,

[0153] Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and They represent the taper of the second half of the Gaussian curve and the taper of the first half of the Gaussian curve, respectively, and f represents the frequency.

[0154] In some embodiments, the training process of the neural network recognition model in the output module 530 is:

[0155] a collecting submodule configured to collect current signal sample data, wherein the current signal sample data includes current signal sample data of multiple low-voltage loads in normal and fault states and current signal sample data of a state in which at least two low-voltage loads are turned on and working simultaneously;

[0156] an expansion submodule, configured to expand the current signal sample data by adopting a cyclic noise addition method to determine the expanded current signal sample data;

[0157] A division submodule is configured to randomly divide the expanded current signal sample data into a training set and a test set according to a preset ratio;

[0158] The training submodule is configured to train the neural network recognition model based on the training set and the test set.

[0159] Exemplary electronic devices

[0160] FIG6 shows the structure of an electronic device provided by an embodiment of the present disclosure. As shown in FIG6 , the electronic device 60 includes one or more processors 61 and a memory 62 .

[0161] The processor 61 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0162] The memory 62 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory (cache), etc. The non-volatile memory may, for example, include read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 61 may execute the program instructions to implement the methods of the software programs of the various embodiments of the present disclosure described above and / or other desired functions. In one example, the electronic device may further include: an input device 63 and an output device 64, which are interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0163] In addition, the input device 63 may also include, for example, a keyboard, a mouse, etc.

[0164] The output device 64 can output various information to the outside. The output device 64 can include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto.

[0165] Of course, for simplicity, FIG6 only shows some of the components of the electronic device related to the present disclosure, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device may further include any other appropriate components depending on the specific application.

[0166] Exemplary computer program products and computer-readable storage media

[0167] In addition to the above-mentioned methods and devices, an embodiment of the present disclosure may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to perform the steps of the method according to various embodiments of the present disclosure described in the above-mentioned "Exemplary Method" section of this specification.

[0168] The computer program product may be written in any combination of one or more programming languages ​​to implement the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, and conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0169] In addition, an embodiment of the present disclosure may also be a computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, enable the processor to execute the steps of the method according to various embodiments of the present disclosure described in the above “Exemplary Method” section of this specification.

[0170] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to a system, system or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0171] The basic principles of the present disclosure have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, and effects mentioned in this disclosure are merely illustrative and not restrictive, and should not be construed as necessarily possessed by each embodiment of the present disclosure. Furthermore, the specific details disclosed above are provided for illustrative purposes and to facilitate understanding, rather than as limitations. These details do not limit the present disclosure to necessarily being implemented using these specific details.

[0172] Each embodiment in this specification is described in a progressive manner, with each embodiment focusing on its differences from the other embodiments. References to the same or similar parts between the various embodiments are sufficient. For system embodiments, since they largely correspond to method embodiments, their description is relatively simple. For relevant parts, references to the description of the method embodiments are sufficient.

[0173] The block diagrams of the devices, systems, equipment, and systems involved in this disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As will be appreciated by those skilled in the art, these devices, systems, equipment, and systems can be connected, arranged, or configured in any manner. Words such as "include," "comprise," "have," and the like are open-ended words, meaning "including but not limited to," and can be used interchangeably therewith. The words "or" and "and" used herein refer to the words "and / or" and can be used interchangeably therewith, unless the context clearly indicates otherwise. The word "such as" used herein refers to the phrase "such as but not limited to," and can be used interchangeably therewith.

[0174] The methods and systems of the present disclosure may be implemented in many ways. For example, the methods and systems of the present disclosure may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is for illustration only, and the steps of the method of the present disclosure are not limited to the order specifically described above unless otherwise specified. In addition, in some embodiments, the present disclosure may also be implemented as programs recorded in a recording medium, which include machine-readable instructions for implementing the methods according to the present disclosure. Thus, the present disclosure also covers recording media that store programs for executing the methods according to the present disclosure.

[0175] It should also be noted that, in the systems, devices and methods of the present disclosure, each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present disclosure. The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present disclosure. Various modifications to these aspects will be very obvious to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the present disclosure. Therefore, the present disclosure is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0176] The above description has been provided for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the present disclosure to the forms disclosed herein. Although a number of example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof. Industrial Applicability

[0177] The embodiments of the present disclosure disclose a method, device, medium and equipment for identifying low-voltage fault arcs. The method includes: using the periodic difference method to extract features of the low-frequency signal within the 3kHz range of the collected electric energy meter, and determining the time domain feature vector of the low-frequency fault arc; using the generalized S transform to extract features of the medium-frequency current signal within 100kHz of the collected electric energy meter, and determining the energy characteristics of the medium-frequency fault arc; inputting the time domain feature vector of the low-frequency fault arc and the energy characteristics of the medium-frequency fault arc into a pre-trained neural network recognition model, and outputting the information on the existence of the fault arc signal; and determining whether the fault arc signal exists based on the information on the existence of the fault arc signal. The above scheme identifies the fault arc when a fault arc occurs in the customer-side circuit, and can perform effective measurement and diagnosis to prevent electrical fire accidents, providing technical support for protecting electricity safety, maintaining the order of electricity supply and use, and ensuring the safe and stable operation of the power grid.

Claims

1. A low voltage fault arc identification method, comprising: The period difference method is used to extract the features of the low-frequency signal within the 3kHz range of the collected electric energy meter to determine the time domain feature vector of the low-frequency fault arc; The generalized S transform is used to extract the features of the 100kHz intermediate frequency current signal collected from the electric energy meter to determine the energy characteristics of the intermediate frequency fault arc. Inputting the low-frequency fault arc time domain feature vector and the medium-frequency fault arc energy feature into a pre-trained neural network recognition model, and outputting the existence information of the fault arc signal; According to the arc fault signal existence information, determining whether there is an arc fault signal; wherein, The generalized S transform is used to extract the features of the 100kHz intermediate frequency current signal collected from the electric energy meter to determine the energy characteristics of the intermediate frequency fault arc, including: Collect the intermediate frequency current signal within 100kHz of the electric energy meter and perform noise reduction and filtering processing; Performing harmonic analysis on the intermediate frequency current signal to determine the harmonic content; For the intermediate frequency current signal whose harmonic content is greater than a preset percentage, a hyperbolic Gaussian window function is used to perform harmonic elimination processing; For the intermediate frequency current signal whose harmonic content is not greater than the preset percentage, a Gaussian window function is used to perform harmonic elimination processing; The generalized S transform is used to extract the current characteristics of the intermediate frequency current signal after the harmonic elimination process, so as to determine the energy characteristics of the intermediate frequency fault arc; The hyperbolic Gaussian window function is: When t ≥ τ, When t<τ, Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and They represent the taper of the second half of the Gaussian curve and the taper of the first half of the Gaussian curve respectively, and f represents the frequency.

2. The method according to claim 1, wherein: The period difference method is used to extract the features of the low-frequency signal within the 3kHz range of the collected electric energy meter to determine the time domain feature vector of the low-frequency fault arc, including: Collect low-frequency signals within the 3kHz range of the electric energy meter, wherein the low-frequency signals include: low-frequency current, voltage and event information; Calculating the low-frequency power consumption characteristics of any period of the low-frequency signal to form a low-frequency time-domain feature vector of the low-frequency signal; The difference between adjacent period feature vectors is compared based on the low-frequency time-domain feature vector to form the low-frequency fault arc time-domain feature vector.

3. The method according to claim 1 or 2, wherein: Comparing the differences of adjacent period feature vectors according to the low-frequency time domain feature vector to form the low-frequency fault arc time domain feature vector includes: Determining signal similarity of a plurality of periodic signals according to calculated current differences between adjacent periods; Normalizing the signal similarity to determine normalized signal similarity; Calculating a period difference according to the signal similarity and the normalized signal similarity; The time domain characteristic vector of the low-frequency fault arc is determined according to the period difference.

4. The method according to claim 1, wherein: The training process of the neural network recognition model is: Collecting current signal sample data, wherein the current signal sample data includes current signal sample data of a plurality of low-voltage loads in normal and fault states and current signal sample data of a state in which at least two low-voltage loads are turned on and working simultaneously; Expanding the current signal sample data by a cyclic noise addition method to determine expanded current signal sample data; The expanded current signal sample data is randomly divided into a training set and a test set according to a preset ratio; The neural network recognition model is trained according to the training set and the test set.

5. A low voltage fault arc identification device, comprising: The first extraction module is configured to extract features from the low-frequency signal within the 3kHz range of the collected electric energy meter using a period difference method to determine a time domain feature vector of the low-frequency fault arc; The second extraction module is configured to use a generalized S transform to extract features from the collected intermediate frequency current signal within 100kHz of the electric energy meter to determine the energy features of the intermediate frequency fault arc; An output module is configured to input the low-frequency fault arc time domain feature vector and the medium-frequency fault arc energy feature into a pre-trained neural network recognition model, and output the fault arc signal existence information; A determination module, configured to determine whether there is a fault arc signal according to the fault arc signal existence information; The second extraction module includes: Collect the intermediate frequency current signal within 100kHz of the electric energy meter and perform noise reduction and filtering processing; Performing harmonic analysis on the intermediate frequency current signal to determine the harmonic content; For the intermediate frequency current signal whose harmonic content is greater than a preset percentage, a hyperbolic Gaussian window function is used to perform harmonic elimination processing; For the intermediate frequency current signal whose harmonic content is not greater than the preset percentage, a Gaussian window function is used to perform harmonic elimination processing; The generalized S transform is used to extract the current characteristics of the intermediate frequency current signal after the harmonic elimination process, so as to determine the energy characteristics of the intermediate frequency fault arc; The hyperbolic Gaussian window function is: When t ≥ τ, When t<τ, Among them, γ bi represents the number of Fourier sine cycles contained within one standard deviation of the Gaussian window, w bi The taper rate is a function of (τ-t), expressed as Indicates, t represents time, τ represents the translation position of the modulation window function along the time axis, and They represent the taper of the second half of the Gaussian curve and the taper of the first half of the Gaussian curve respectively, and f represents the frequency.

6. A computer-readable storage medium, wherein: The storage medium stores a computer program, and the computer program is configured to execute the method according to any one of claims 1 to 4.

7. An electronic device, wherein: The electronic device comprises: processor; a memory configured to store instructions executable by the processor; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1 to 4.

Images