Fault classification method and device for arc current signal, and storage medium

Abstract

The invention discloses a fault classification method and device for arc current signals and a storage medium. The method comprises the following steps: collecting a current arc current signal, and preprocessing the current arc current signal to obtain a to-be-decomposed arc current signal; decomposing the arc current signal to be decomposed into a plurality of single fluctuation components based on a local fluctuation feature decomposition method, performing multi-dimensional feature data extraction on each single fluctuation component, and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; and determining a fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model. Accurate fault identification of the arc current signal is realized, and the efficiency and accuracy of fault classification are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of current signal analysis, and in particular to a fault classification method, device and storage medium for arc current signals. Background Art

[0002] The power system is a vital infrastructure in modern society. Its reliability and stability directly affect the normal operation of society. Arc faults are a common and dangerous type of fault in power systems, which can cause serious equipment damage and safety hazards. Therefore, the detection and identification of arc faults has become an important task in power system maintenance.

[0003] Traditional arc fault detection methods mainly rely on manual inspections and regular maintenance, but these methods have problems such as low efficiency and poor real-time performance. Arc fault signals have complex time-frequency characteristics, and there are still many challenges in fault identification of arc current signals. Summary of the invention

[0004] The present invention provides a method, a device and a storage medium for fault classification of arc current signals, so as to solve the problem of low efficiency in fault identification of arc current signals.

[0005] According to one aspect of the present invention, a method for fault classification of an arc current signal is provided, the method comprising:

[0006] Collecting a current arc current signal, and preprocessing the current arc current signal to obtain an arc current signal to be decomposed;

[0007] Decomposing the arc current signal to be decomposed into multiple single fluctuation components based on a local fluctuation feature decomposition method, extracting multi-dimensional feature data of each single fluctuation component, and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features;

[0008] A fault classification result corresponding to the current arc current signal is determined based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model integrating multiple decision trees.

[0009] According to another aspect of the present invention, there is provided a fault classification device for an arc current signal, the device comprising:

[0010] A signal acquisition module, used for acquiring a current arc current signal and preprocessing the current arc current signal to obtain an arc current signal to be decomposed;

[0011] A feature extraction module, used for decomposing the arc current signal to be decomposed into multiple single fluctuation components based on a local fluctuation feature decomposition method, extracting multi-dimensional feature data of each single fluctuation component, and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features;

[0012] A fault classification module is used to determine a fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model that integrates multiple decision trees.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the arc current signal fault classification method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the arc current signal fault classification method described in any embodiment of the present invention when executed.

[0018] The technical solution of the embodiment of the present invention collects the current arc current signal, pre-processes the current arc current signal to obtain the arc current signal to be decomposed; realizes real-time detection of the arc current signal; then, decomposes the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method, extracts multi-dimensional feature data for each single fluctuation component, and constructs a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; efficiently decomposes the current arc current signal and reconstructs the target feature vector; finally, determines the fault classification result corresponding to the current arc current signal based on the target feature vector corresponding to the multiple single fluctuation components and the pre-trained target signal classification model, wherein the target signal classification model is obtained by training the initial signal classification model integrating multiple decision trees. The problem of low efficiency in fault identification of arc current signals is solved, accurate fault identification of arc current signals is realized, and the beneficial effect of improving the efficiency and accuracy of fault classification is achieved.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 is a flow chart of a fault classification method of an arc current signal provided according to Embodiment 1 of the present invention;

[0022] Figure 2a is a flow chart of a fault classification method of an arc current signal provided according to Embodiment 2 of the present invention;

[0023] Figure 2b is a flow chart of an optional example of a fault classification method of an arc current signal provided in accordance with the second embodiment of the present invention;

[0024] Figure 3 2 is a schematic diagram of the structure of a fault classification device for arc current signals provided in accordance with Embodiment 3 of the present invention;

[0025] Figure 4 It is a structural schematic diagram of an electronic device for implementing the fault classification method of arc current signal according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] Embodiment 1

[0029] Figure 1 A flowchart of a method for fault classification of arc current signals is provided for the first embodiment of the present invention. This embodiment is applicable to the case of current signal analysis. The method can be performed by a fault classification device for arc current signals. The fault classification device for arc current signals can be implemented in the form of hardware and / or software. The fault classification device for arc current signals can be configured in an electronic device. Figure 1 As shown, the method includes:

[0030] S110, collecting a current arc current signal, and preprocessing the current arc current signal to obtain an arc current signal to be decomposed.

[0031] The current arc current signal may be understood as a signal formed by the current change generated during the arc discharge process acquired in real time.

[0032] Specifically, the arc current signal is collected by a current sensor with a high sampling rate, and the collected data is preprocessed, and the preprocessed arc current signal is used as the arc current signal to be decomposed.

[0033] Optionally, preprocessing the current arc current signal to obtain the arc current signal to be decomposed includes:

[0034] The current arc current signal is subjected to denoising, normalization and data segmentation processing to obtain the arc current signal to be decomposed.

[0035] Specifically, the current arc current signal is denoised, and a filter is used to remove high-frequency noise and interference in the signal to ensure the purity of the signal; the current arc current signal is normalized, and signals of different amplitudes are adjusted to a uniform range to eliminate the influence of amplitude differences on feature extraction and model training; the current arc current signal is data segmented, and the preprocessed signal data is segmented according to time windows to generate multiple data segments.

[0036] Specifically, a filter is used to remove high-frequency noise and interference in the signal. The selection of the filter can be determined according to the specific application requirements and noise characteristics. Considering the analysis of the current of household appliances, the filter is selected as a low-pass Butterworth filter. The filter is expressed by the following formula:

[0037]

[0038] Where S is the complex frequency, w c is the cutoff frequency and n is the filter order.

[0039] w c =2πf c ;

[0040] Among them, f c is the cut-off frequency, w c is the angular frequency.

[0041] The normalization process of the current arc current signal is expressed as follows:

[0042]

[0043] Among them, x' is the normalized data, x is the original data, and x min is the minimum value of the original data, x max is the maximum value of the original data.

[0044] The data segmentation process of the current arc current signal is expressed by the following formula:

[0045] x i =x(t+(il)L),i=1,2,3…N;

[0046] Where N is the number of data segments after segmentation, and t is the initial time.

[0047] In the embodiment of the present invention, by performing denoising, normalization and data segmentation processing on the current arc current signal, the signal quality and analysis accuracy are significantly improved, which facilitates subsequent signal processing.

[0048] S120. Decompose the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method, extract multi-dimensional feature data of each single fluctuation component, and construct a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features.

[0049] The single fluctuation component can be understood as the fluctuation component in the arc current signal to be decomposed. The arc current signal may contain multiple fluctuation components with different frequencies and amplitudes, each of which represents a certain physical phenomenon or process in the arc discharge process.

[0050] Specifically, the arc current signal to be decomposed is subjected to piecewise linear transformation and rational spline fitting to obtain multiple single fluctuation components. Multi-dimensional feature data is extracted from the single fluctuation component, and a target feature vector corresponding to the single fluctuation component is constructed based on the extracted multi-dimensional features.

[0051] Optionally, the local fluctuation feature decomposition method decomposes the arc current signal to be decomposed into multiple single fluctuation components, including: determining the decomposition scale of the arc current signal to be decomposed, and determining the decomposition level of the arc current signal to be decomposed based on the decomposition scale; performing piecewise linear transformation on the arc current signal to be decomposed at each decomposition level to obtain multiple linear segments; and smoothing each of the linear segments based on a preset fitting method to obtain multiple single fluctuation components.

[0052] Among them, the decomposition scale can be understood as the degree of fineness of signal decomposition.

[0053] Specifically, the decomposition scale is selected according to the characteristics of the arc current signal and the actual application requirements. Among them, a smaller decomposition scale can capture the detailed changes of the signal, while a larger decomposition scale is suitable for extracting the overall trend of the signal. The arc current signal to be decomposed is transformed piecewise linearly. The arc current signal to be decomposed is locally linearly fitted to convert the complex nonlinear signal into several linear segments, thereby simplifying subsequent processing. Each linear segment represents the change trend of the signal within the interval, which is helpful for subsequent component extraction. Each linear segment is smoothed using the rational spline fitting method. It can eliminate noise and mutation points in the signal, improve the smoothness and stability of the decomposition result; generate multiple smoothed single fluctuation components, each of which represents the characteristics of the signal at a specific frequency and amplitude. The smoothed fluctuation component is feature extracted to construct a feature vector.

[0054] For example, the arc current signal sampling rate is 100kHz, and the analysis frequency range is 1-10kHz. The initial decomposition scale S=1KHZ is set, which means that the time length of each segment is 1 millisecond. The pre-processed arc current signal to be decomposed is subjected to piecewise linear transformation, and the complex nonlinear signal is converted into several linear segments by performing local linear fitting on the signal. For each time period [t i ,t i+1 ], the signal can be expressed as a linear function using the following formula:

[0055] x(t)≈a i t+b i t(t∈[t i ,t i+1 ]);

[0056] Among them, a i and b i are the fitting parameters over this time period.

[0057] Each linear segment is smoothed using a rational spline fitting method.

[0058] The rational spline function is expressed by the following formula:

[0059]

[0060] Among them, P j (t) is a polynomial function, w j (t) is the weight function and n is the order of the spline.

[0061] In the embodiment of the present invention, the local fluctuation feature decomposition method can efficiently decompose the arc current signal to be decomposed into several representative fluctuation components, laying a solid foundation for subsequent feature extraction and fault identification.

[0062] S130. Determine a fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model integrating multiple decision trees.

[0063] Among them, the fault classification results include series arc fault, parallel arc fault, parallel arc fault and compound arc fault.

[0064] Specifically, the target feature vector is input into a pre-trained target signal classification model, and the model outputs one or more possible fault categories as classification results based on the input feature vector.

[0065] Optionally, before determining the fault classification result corresponding to the current arc current signal based on multiple target feature vectors and a pre-trained target signal classification model, it also includes: obtaining a sample data set, dividing the sample data set into a training set, a validation set and a test set, wherein the sample data set includes sample feature vectors under multiple fault categories; constructing an initial signal classification model integrating multiple decision trees, and iteratively training the initial signal classification model based on the training set and a gradient boosting algorithm; during the training process, cross-validating the model after each training based on the validation set, adjusting the model parameters based on the validation results, and evaluating the model after each training based on the test set to obtain a model classification accuracy index; and determining the target signal classification model based on at least one of the model classification accuracy indicators.

[0066] Specifically, the feature vector of the training set is used as the input data of the initial signal classification model, and the model is trained by the gradient boosting algorithm. During the training process, the model continuously adjusts parameters to minimize the error between the predicted result and the actual fault category. In order to evaluate the performance of the model, cross-validation is performed to evaluate the generalization ability and robustness of the model. According to the results of cross-validation, the hyperparameters of the model are adjusted to further optimize the performance of the model to ensure that the model can accurately identify arc faults. The model after each training is evaluated based on the test set to obtain the model classification accuracy index; the model with the model classification accuracy index is determined as the target signal classification model.

[0067] Optionally, the feature vector of the training set is used as the model input data to train the model through the gradient boosting algorithm. The training goal is to minimize the error between the predicted result and the actual fault category. The loss function is expressed as follows:

[0068]

[0069] Among them, L is the loss function, N is the number of samples, and y i is the true label, is the predicted value, and l is the loss function of a single sample.

[0070] Cross validation is performed using the following formula:

[0071]

[0072] Among them, k is the number of cross-validation folds, n i is the number of validation samples for the i-th fold, and are the predicted value and true value of the jth sample in the i-th fold, respectively, and l is the indicator function.

[0073] Furthermore, according to the results of cross-validation, the model's hyperparameters are adjusted to further optimize the model's performance and ensure that the model can accurately identify arc faults. The hyperparameters are adjusted to adjust the tree's depth, learning rate, and number of iterations. Optimization goal: maximize the accuracy of cross-validation and minimize the validation error. The hyperparameter optimization formula is as follows:

[0074]

[0075] Among them, θ is the set of hyperparameters of the model, represents the predicted value of the jth sample in the i-th fold under the hyperparameter θ.

[0076] Optionally, collect new data and retrain and optimize the target signal classification model to improve the accuracy and reliability of fault identification. Collect arc current signals in real time, add new data to the training data set, continuously expand data samples, and improve the generalization ability of the model. Use the updated training data set to retrain the target signal classification model. Improve the recognition performance and stability of the model through continuous iterative training and optimization. Provide timely feedback and adjust model parameters based on the recognition results in actual applications to ensure that the system can maintain efficient and accurate fault recognition capabilities under different working conditions.

[0077] The technical solution of the embodiment of the present invention collects the current arc current signal, pre-processes the current arc current signal to obtain the arc current signal to be decomposed; realizes real-time detection of the arc current signal; then, decomposes the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method, extracts multi-dimensional feature data for each single fluctuation component, and constructs a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; efficiently decomposes the current arc current signal and reconstructs the target feature vector; finally, determines the fault classification result corresponding to the current arc current signal based on the target feature vector corresponding to the multiple single fluctuation components and the pre-trained target signal classification model, wherein the target signal classification model is obtained by training the initial signal classification model integrating multiple decision trees. The problem of low efficiency in fault identification of arc current signals is solved, accurate fault identification of arc current signals is realized, and the beneficial effect of improving the efficiency and accuracy of fault classification is achieved.

[0078] Embodiment 2

[0079] Figure 2aA flowchart of a fault classification method for an arc current signal provided in Embodiment 2 of the present invention, this embodiment is a further optimization of how to extract multi-dimensional feature data of the single fluctuation component in the above embodiment, and construct a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features. Optionally, the multi-dimensional feature data extraction of the single fluctuation component and the construction of a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features include: determining the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component, combining the time domain feature data, the frequency domain feature data and the time-frequency feature data corresponding to the single fluctuation component to construct a target feature vector corresponding to the single fluctuation component.

[0080] like Figure 2a As shown, the method includes:

[0081] S210, collecting a current arc current signal, and preprocessing the current arc current signal to obtain an arc current signal to be decomposed.

[0082] S220. Decompose the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method. For each of the single fluctuation components, determine the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component. Combine the time domain feature data, the frequency domain feature data and the time-frequency feature data corresponding to the single fluctuation component to construct a target feature vector corresponding to the single fluctuation component.

[0083] Specifically, for each of the single fluctuation components, the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component are extracted respectively, and the time domain feature data, frequency domain feature data and time-frequency feature data are combined into a target feature vector. The target feature vector is used to characterize various aspects of the arc current signal to be decomposed, contains information of the signal in different domains, and provides comprehensive feature data for subsequent model training.

[0084] Optionally, the time-frequency characteristic data includes mean, variance, peak value and root mean square value; the time domain characteristic data of the single fluctuation component is determined by the following formula:

[0085]

[0086]

[0087] P = max(x i );

[0088]

[0089] Among them, μ is the mean, σ 2 is the variance, P is the peak value, RMS is the root mean square value, x i is the ith value in a single fluctuation component, and N is the total number of values ​​in a single fluctuation component.

[0090] Optionally, the determining the frequency domain characteristic data of the single fluctuation component includes:

[0091] Each of the single fluctuation components is subjected to Fourier transformation to obtain frequency domain characteristic data corresponding to the single fluctuation component.

[0092] Among them, the frequency domain feature data includes main frequency component, spectrum energy and bandwidth.

[0093] Optionally, each of the single fluctuation components is Fourier transformed using the following formula:

[0094]

[0095] Main frequency component: main frequency f peak It is the frequency component with the largest amplitude in the spectrum.

[0096] Spectral energy:

[0097] Bandwidth: The range over which energy in the frequency spectrum is distributed.

[0098] Optionally, the time-frequency characteristic data of the single fluctuation component is determined by the following formula, including:

[0099]

[0100] Among them, W x (a, b) are wavelet coefficients, x(t) is a single wave component, a is a scale parameter, b is a translation parameter, and Ψ is the mother wavelet function.

[0101] Optionally, the time domain feature data, the frequency domain feature data and the time-frequency feature data are combined into a target feature vector for characterizing various aspects of the arc current signal:

[0102] F=[μ,σ 2 ,P,RMS,f peak ,E,BW,…,];

[0103] Through the above steps, the arc current signal is decomposed into multiple representative fluctuating components, each of which contains rich signal features, laying a solid foundation for subsequent feature extraction and fault identification.

[0104] S230. Determine a fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model that integrates multiple decision trees.

[0105] The technical solution of the embodiment of the present invention determines the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component, and combines the time domain feature data, frequency domain feature data and time-frequency feature data corresponding to the single fluctuation component to construct a target feature vector corresponding to the single fluctuation component. The target feature vector contains information about the signal in different domains, providing comprehensive feature data for subsequent model training.

[0106] Figure 2b A flow chart of an optional example of a fault classification method for arc current signals is provided. Figure 2b As shown, the arc current signal fault classification method specifically includes the following steps:

[0107] Step 1) data collection and preprocessing;

[0108] Data acquisition: Use a current sensor with a high sampling rate (e.g. 100kHz) to collect arc current signals. The high sampling rate ensures that the high-frequency details of the signal are captured, thereby improving the accuracy of subsequent signal analysis.

[0109] Use filters to remove high-frequency noise and interference in the signal. The choice of filter can be determined according to the specific application requirements and noise characteristics. Considering the analysis of household appliance current, the filter is selected as a low-pass filter. The filter is expressed by the following formula:

[0110]

[0111] Where S is the complex frequency, w c is the cutoff frequency and n is the filter order.

[0112] w c =2πf c ;

[0113] Among them, f c is the cut-off frequency, w c is the angular frequency.

[0114] The normalization process of the current arc current signal is expressed as follows:

[0115]

[0116] Among them, x' is the normalized data, x is the original data, and x min is the minimum value of the original data, x max is the maximum value of the original data.

[0117] The data segmentation process of the current arc current signal is expressed by the following formula:

[0118] x i =x(t+(il)L),i=1,2,3…N;

[0119] Where N is the number of data segments after segmentation, and t is the initial time.

[0120] In the embodiment of the present invention, by performing denoising, normalization and data segmentation processing on the current arc current signal, the signal quality and analysis accuracy are significantly improved, which facilitates subsequent signal processing.

[0121] Step 2) Decompose the signal using the local fluctuation feature decomposition method:

[0122] Choose the appropriate decomposition scale. The decomposition scale determines the fineness of the signal decomposition. Smaller decomposition scale: can capture the detailed changes of the signal. Larger decomposition scale: suitable for extracting the overall trend of the signal. The signal x(t) decomposition scale S determines the decomposition level of the signal. Now the arc current signal sampling rate is 100kHz, and the analysis frequency range is 1-10kHz. Set the initial decomposition scale S = 1KHZ, which means that the length of each segment time is 1 millisecond.

[0123] After selecting the decomposition scale, perform piecewise linear transformation: perform piecewise linear transformation on the preprocessed arc current signal to be decomposed, and convert the complex nonlinear signal into several linear segments by performing local linear fitting on the signal. i ,t i+1 ], the signal can be expressed as a linear function using the following formula:

[0124] x(t)≈a i t+b i t(t∈[t i ,t i+1 ])

[0125] Among them, a i and b i are the fitting parameters over this time period.

[0126] Each linear segment is smoothed using a rational spline fitting method.

[0127] The rational spline function is expressed by the following formula:

[0128]

[0129] Among them, P j (t) is a polynomial function, w j (t) is the weight function and n is the order of the spline.

[0130] In the embodiment of the present invention, the local fluctuation feature decomposition method can efficiently decompose the arc current signal into several representative fluctuation components, laying a solid foundation for subsequent feature extraction and fault identification.

[0131] Step 3) Feature extraction: Extract features of the single fluctuation component obtained by decomposition: Extract features of the single fluctuation component after smoothing and construct a feature vector.

[0132] Time domain characteristics: The time-frequency characteristic data includes mean, variance, peak value and root mean square value; the time domain characteristic data of the single fluctuation component is determined by the following formula:

[0133]

[0134]

[0135] P = max(x i );

[0136]

[0137] Among them, μ is the mean, σ 2 is the variance, P is the peak value, RMS is the root mean square value, x i is the ith value in a single fluctuation component, and N is the total number of values ​​in a single fluctuation component.

[0138] Optionally, the determining the frequency domain characteristic data of the single fluctuation component includes:

[0139] Each of the single fluctuation components is subjected to Fourier transformation to obtain frequency domain characteristic data corresponding to the single fluctuation component.

[0140] Among them, the frequency domain feature data includes main frequency component, spectrum energy and bandwidth.

[0141] Optionally, each of the single fluctuation components is Fourier transformed using the following formula:

[0142]

[0143] Main frequency component: main frequency f peak It is the frequency component with the largest amplitude in the spectrum.

[0144] Spectral energy:

[0145] Bandwidth: The range over which energy in the frequency spectrum is distributed.

[0146] Optionally, the time-frequency characteristic data of the single fluctuation component is determined by the following formula, including:

[0147]

[0148] Among them, W x (a, b) are wavelet coefficients, x(t) is a single wave component, a is a scale parameter, b is a translation parameter, and Ψ is the mother wavelet function.

[0149] Feature vector construction: The time domain, frequency domain and time-frequency features are combined into a feature vector F, which is used to characterize various aspects of the arc current signal.

[0150] F=[μ,σ 2 ,P,RMS,f peak ,E,BW,…,]

[0151] Through the above steps, the arc current signal is decomposed into multiple representative fluctuating components, each of which contains rich signal features, laying a solid foundation for subsequent feature extraction and fault identification.

[0152] Step 4) Model training

[0153] Model initialization

[0154] A sample data set is obtained, and the sample data set is divided into a training set, a validation set, and a test set, wherein the sample data set includes sample feature vectors under multiple fault categories; an initial signal classification model integrating multiple decision trees is constructed, and the initial signal classification model is iteratively trained based on the training set and a gradient boosting algorithm; during the training process, the model after each training is cross-validated based on the validation set, the model parameters are adjusted based on the validation results, and the model after each training is evaluated based on the test set to obtain a model classification accuracy index; and a target signal classification model is determined based on at least one of the model classification accuracy indicators.

[0155] Specifically, the feature vector of the training set is used as the input data of the initial signal classification model, and the model is trained by the gradient boosting algorithm. During the training process, the model continuously adjusts parameters to minimize the error between the predicted result and the actual fault category. In order to evaluate the performance of the model, cross-validation is performed to evaluate the generalization ability and robustness of the model. According to the results of cross-validation, the hyperparameters of the model are adjusted to further optimize the performance of the model to ensure that the model can accurately identify arc faults. The model after each training is evaluated based on the test set to obtain the model classification accuracy index; the model with the model classification accuracy index is determined as the target signal classification model.

[0156] Optionally, the feature vector of the training set is used as the model input data to train the model through the gradient boosting algorithm. The training goal is to minimize the error between the predicted result and the actual fault category. The loss function is expressed as follows:

[0157]

[0158] Among them, L is the loss function, n is the number of samples, y i is the true label, is the predicted value, and l is the loss function of a single sample.

[0159] K-fold cross validation:

[0160] Divide the data set into K subsets, use K-1 subsets for training each time, and use the remaining subset for validation, repeating K times. Calculate the accuracy of each validation and take the average as the performance indicator of the model. Perform cross validation using the following formula:

[0161]

[0162] Among them, k is the number of cross-validation folds, n i is the number of validation samples for the i-th fold, and are the predicted value and true value of the jth sample in the i-th fold, respectively, and l is the indicator function.

[0163] Furthermore, according to the results of cross-validation, the model's hyperparameters are adjusted to further optimize the model's performance and ensure that the model can accurately identify arc faults. The hyperparameters are adjusted to adjust the tree's depth, learning rate, and number of iterations. Optimization goal: maximize the accuracy of cross-validation and minimize the validation error. The hyperparameter optimization formula is as follows:

[0164]

[0165] Among them, θ is the set of hyperparameters of the model, represents the predicted value of the jth sample in the i-th fold under the hyperparameter θ.

[0166] Optionally, collect new data and retrain and optimize the target signal classification model to improve the accuracy and reliability of fault identification. Collect arc current signals in real time, add new data to the training data set, continuously expand data samples, and improve the generalization ability of the model. Use the updated training data set to retrain the target signal classification model. Improve the recognition performance and stability of the model through continuous iterative training and optimization. Provide timely feedback and adjust model parameters based on the recognition results in actual applications to ensure that the system can maintain efficient and accurate fault recognition capabilities under different working conditions.

[0167] Step 5) Fault identification and model optimization: Fault classification: The newly collected arc current signal is

[0168] After preprocessing and feature extraction, it is input into the trained target signal classification model for fault classification. Model optimization: According to the results of fault classification, analyze the recognition accuracy and error distribution of the model. For misclassified samples, re-extract features and retrain the model to continuously optimize the performance of the model and improve the accuracy of fault identification. Specific steps:

[0169] Analyze classification results: Count classification results, calculate accuracy, and analyze error distribution.

[0170] Identify misclassified samples: Identify misclassified samples and analyze their feature distribution.

[0171] Feature extraction optimization: Re-extract the features of misclassified samples and check for problems in the feature extraction process.

[0172] Model retraining: Re-input the optimized feature vector into the model for training and update the model parameters.

[0173] Iterative optimization: Repeat the above steps to continuously optimize the performance of the model.

[0174] Through the above specific implementation steps, the trained target signal classification model is used to classify and identify new arc current signals, and the performance of the model is continuously optimized according to the classification results to ensure the accuracy and robustness of fault identification.

[0175] Step 6) System feedback and update: Collect new data and retrain and optimize the target signal classification model to improve the accuracy and reliability of fault identification. Collect arc current signals in real time, add new data to the training data set, continuously expand data samples, and improve the generalization ability of the model. Use the updated training data set to retrain the target signal classification model. Improve the recognition performance and stability of the model through continuous iterative training and optimization. According to the recognition results in actual applications, timely feedback and adjustment of model parameters are provided to ensure that the system can maintain efficient and accurate fault recognition capabilities under different working conditions.

[0176] The technical solution of the embodiment of the present invention can adaptively decompose the signal through the local fluctuation feature decomposition method, decompose the complex arc signal into multiple single fluctuation components, so as to better extract the fault characteristics in the signal. Combined with the template signal classification model, the gradient boosting algorithm is used for efficient feature learning and fault identification, which effectively improves the accuracy and real-time performance of arc fault detection, which can not only improve the fault monitoring capability of power equipment, but also provide strong guarantee for the safe operation of the power system.

[0177] Embodiment 3

[0178] Figure 3 This is a schematic diagram of the structure of a fault classification device for arc current signals provided in the third embodiment of the present invention. Figure 3 As shown, the device includes: a signal acquisition module 310 , a feature extraction module 320 and a fault classification module 330 .

[0179] Among them, the signal acquisition module 310 is used to collect the current arc current signal and pre-process the current arc current signal to obtain the arc current signal to be decomposed; the feature extraction module 320 is used to decompose the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method, and for each single fluctuation component, multi-dimensional feature data is extracted for the single fluctuation component, and a target feature vector corresponding to the single fluctuation component is constructed based on the extracted multi-dimensional features; the fault classification module 330 is used to determine the fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model that integrates multiple decision trees.

[0180] The technical solution of the embodiment of the present invention is to collect the current arc current signal through the signal acquisition module, pre-process the current arc current signal to obtain the arc current signal to be decomposed; realize the real-time detection of the arc current signal; then, through the feature extraction module, decompose the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method, extract multi-dimensional feature data for each single fluctuation component, and construct the target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; efficiently decompose the current arc current signal and reconstruct the target feature vector; finally, through the fault classification module, determine the fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and the pre-trained target signal classification model, wherein the target signal classification model is obtained by training the initial signal classification model integrating multiple decision trees. The problem of low efficiency in fault identification of arc current signals is solved, accurate fault identification of arc current signals is realized, and the beneficial effect of improving the efficiency and accuracy of fault classification is achieved.

[0181] Optionally, the feature extraction module includes:

[0182] A decomposition level determination unit, used to determine a decomposition scale of the arc current signal to be decomposed, and determine a decomposition level of the arc current signal to be decomposed based on the decomposition scale;

[0183] A linear segment acquisition unit, used for performing piecewise linear transformation on the arc current signal to be decomposed at each decomposition level to obtain a plurality of linear segments;

[0184] The single fluctuation component acquisition unit is used to perform smoothing processing on each of the linear segments based on a preset fitting method to obtain a plurality of the single fluctuation components.

[0185] Optionally, the feature extraction module is specifically used for:

[0186] Determine the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component, and combine the time domain feature data, the frequency domain feature data and the time-frequency feature data corresponding to the single fluctuation component to construct a target feature vector corresponding to the single fluctuation component.

[0187] Optionally, the time-frequency feature data includes mean, variance, peak and root mean square value; accordingly, the feature extraction module is specifically used to: determine the time domain feature data of the single fluctuation component by the following formula:

[0188]

[0189]

[0190] P = max(x i );

[0191]

[0192] Among them, μ is the mean, σ 2 is the variance, P is the peak value, RMS is the root mean square value, x i is the ith value in a single fluctuation component, and N is the total number of values ​​in a single fluctuation component.

[0193] Optionally, the feature extraction module is specifically used for:

[0194] Each of the single fluctuation components is subjected to Fourier transformation to obtain frequency domain characteristic data corresponding to the single fluctuation component.

[0195] Optionally, the feature extraction module is specifically used to determine the time-frequency feature data of the single fluctuation component by the following formula, including:

[0196]

[0197] Among them, W x (a, b) are wavelet coefficients, x(t) is a single wave component, a is a scale parameter, b is a translation parameter, and Ψ is the mother wavelet function.

[0198] Optionally, the device also includes a data set partitioning module, a model training module, a model evaluation module and a target model determination module.

[0199] The data set division module is used to obtain a sample data set and divide the sample data set into a training set, a validation set and a test set before determining the fault classification result corresponding to the current arc current signal based on the multiple target feature vectors and the pre-trained target signal classification model, wherein the sample data set includes sample feature vectors under multiple fault categories;

[0200] The model training module is used to construct an initial signal classification model integrating multiple decision trees, and iteratively train the initial signal classification model based on the training set and the gradient boosting algorithm;

[0201] The model evaluation module is used to perform cross-validation on the model after each training based on the validation set during the training process, adjust the model parameters based on the validation results, and evaluate the model after each training based on the test set to obtain a model classification accuracy index;

[0202] The target model determination module is used to determine the target signal classification model based on at least one of the model classification accuracy indicators.

[0203] Optionally, the signal acquisition module is specifically used for:

[0204] The current arc current signal is subjected to denoising, normalization and data segmentation processing to obtain the arc current signal to be decomposed.

[0205] The arc current signal fault classification device provided in the embodiment of the present invention can execute the arc current signal fault classification method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0206] Embodiment 4

[0207] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0208] like Figure 4 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0209] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0210] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the method for fault classification of arc current signals.

[0211] In some embodiments, the method for fault classification of arc current signals may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for fault classification of arc current signals described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the method for fault classification of arc current signals in any other suitable manner (e.g., by means of firmware).

[0212] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0213] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0214] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer 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 of the foregoing.

[0215] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0216] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0217] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0218] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0219] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A fault classification method for arc current signal, characterized in that: include: Collecting a current arc current signal, and preprocessing the current arc current signal to obtain an arc current signal to be decomposed; Decomposing the arc current signal to be decomposed into multiple single fluctuation components based on a local fluctuation feature decomposition method, extracting multi-dimensional feature data of each single fluctuation component, and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; A fault classification result corresponding to the current arc current signal is determined based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model integrating multiple decision trees.

2. The method according to claim 1, characterized in that The method for decomposing the arc current signal to be decomposed into multiple single fluctuation components based on the local fluctuation feature decomposition method includes: Determining a decomposition scale of the arc current signal to be decomposed, and determining a decomposition level of the arc current signal to be decomposed based on the decomposition scale; Performing piecewise linear transformation on the arc current signal to be decomposed at each decomposition level to obtain a plurality of linear segments; Each of the linear segments is smoothed based on a preset fitting method to obtain a plurality of the single fluctuation components.

3. The method according to claim 1, characterized in that: The step of extracting multi-dimensional feature data from the single fluctuation component and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features includes: Determine the time domain feature data, frequency domain feature data and time-frequency feature data of the single fluctuation component, and combine the time domain feature data, the frequency domain feature data and the time-frequency feature data corresponding to the single fluctuation component to construct a target feature vector corresponding to the single fluctuation component.

4. The method according to claim 3, characterized in that The time-frequency characteristic data includes mean, variance, peak value and root mean square value; the time domain characteristic data of the single fluctuation component is determined by the following formula: P=max(x i ); Among them, μ is the mean, σ 2 is the variance, P is the peak value, RMS is the root mean square value, x i is the ith value in a single fluctuation component, and N is the total number of values ​​in a single fluctuation component.

5. The method according to claim 4, characterized in that The determining of the frequency domain characteristic data of the single fluctuation component comprises: Each of the single fluctuation components is subjected to Fourier transformation to obtain frequency domain characteristic data corresponding to the single fluctuation component.

6. The method according to claim 4, characterized in that The time-frequency characteristic data of the single fluctuation component is determined by the following formula, including: Among them, W x (a, b) are wavelet coefficients, x(t) is a single wave component, a is a scale parameter, b is a translation parameter, and Ψ is the mother wavelet function.

7. The method according to claim 1, characterized in that Before determining the fault classification result corresponding to the current arc current signal based on the plurality of target feature vectors and the pre-trained target signal classification model, the method further includes: Acquire a sample data set, and divide the sample data set into a training set, a validation set, and a test set, wherein the sample data set includes sample feature vectors under multiple fault categories; Constructing an initial signal classification model integrating multiple decision trees, and iteratively training the initial signal classification model based on the training set and the gradient boosting algorithm; During the training process, the model after each training is cross-validated based on the validation set, the model parameters are adjusted based on the validation results, and the model after each training is evaluated based on the test set to obtain a model classification accuracy index; A target signal classification model is determined based on at least one of the model classification accuracy indicators.

8. The method according to claim 1, characterized in that: The preprocessing of the current arc current signal to obtain the arc current signal to be decomposed includes: The current arc current signal is subjected to denoising, normalization and data segmentation processing to obtain an arc current signal to be decomposed.

9. A fault classification device for arc current signals, characterized in that: include: A signal acquisition module, used for acquiring a current arc current signal and preprocessing the current arc current signal to obtain an arc current signal to be decomposed; A feature extraction module, used for decomposing the arc current signal to be decomposed into multiple single fluctuation components based on a local fluctuation feature decomposition method, extracting multi-dimensional feature data of each single fluctuation component, and constructing a target feature vector corresponding to the single fluctuation component based on the extracted multi-dimensional features; A fault classification module is used to determine a fault classification result corresponding to the current arc current signal based on the target feature vectors corresponding to the multiple single fluctuation components and a pre-trained target signal classification model, wherein the target signal classification model is obtained by training an initial signal classification model that integrates multiple decision trees.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the arc current signal fault classification method according to any one of claims 1 to 8 when executed.

Images