Method, device and equipment for identifying short-circuit fault type of power distribution network and medium

Abstract

The invention provides a power distribution network short circuit fault type identification method and device, equipment and a medium, and relates to the technical field of electric power. The method comprises the steps of obtaining time sequence operation data of a bus node corresponding to a fault line in response to a short-circuit fault of a power distribution network; dimension raising processing is carried out on the time sequence operation data through the observation function, and an observation space data matrix is obtained; fitting the observation space data matrix to obtain an equivalent fault response model corresponding to the power distribution network; based on the equivalent fault response model, extracting a target fault feature vector; and identifying a target fault type corresponding to the power distribution network based on the target fault feature vector. The short-circuit fault type of the power distribution network can be identified more accurately and efficiently.

Description

Technical Field

[0001] The present application relates to the technical field of electric power, and in particular, to a method, device, equipment and medium for identifying types of short - circuit faults in a distribution network. Background Art

[0002] As an infrastructure, the safety and stability of the operation of the power system are of crucial importance. Among them, the distribution network is an indispensable part of the power system, responsible for the final distribution and delivery of electric power to ensure that users obtain reliable and stable power supply. Therefore, when a short - circuit fault occurs in the distribution network, quickly and accurately identifying the type of short - circuit fault has important research significance.

[0003] Currently, when a short - circuit fault occurs in the distribution network, a data - driven pattern recognition method is usually adopted to identify the type of short - circuit fault in the distribution network. Among them, machine learning or statistical methods are used to extract data features from the operation data of the distribution network for identifying the type of short - circuit fault. However, when identifying the type of short - circuit fault in the distribution network in the above - mentioned manner, there is a problem that the recognition result is not accurate enough. Summary of the Invention

[0004] Embodiments of the present application provide a method, device, equipment and medium for identifying types of short - circuit faults in a distribution network to solve the problem that the recognition result is not accurate enough when identifying the type of short - circuit fault in the distribution network by the current method.

[0005] In a first aspect, embodiments of the present application provide a method for identifying types of short - circuit faults in a distribution network, including:

[0006] In response to a short - circuit fault occurring in the distribution network, obtaining the time - series operation data of the bus node corresponding to the faulty line;

[0007] Performing dimensionality - raising processing on the time - series operation data through an observation function to obtain an observation space data matrix;

[0008] Performing fitting processing on the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network;

[0009] Based on the equivalent fault response model, extracting a target fault feature vector;

[0010] Based on the target fault feature vector, identifying the target fault type corresponding to the distribution network.

[0011] Optionally, the observation function includes polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions. The time-series operation data is dimensionally elevated through the observation function to obtain an observation space data matrix, including: inputting the time-series operation data into the polynomial basis function, Gaussian kernel function, triangular basis function, and radial basis function respectively to perform dimensional elevation processing on the time-series operation data to obtain the observation space data matrix; wherein, the polynomial basis function is used to obtain the coupling relationship and high-order features between the time-series operation data; the Gaussian kernel function is used to obtain the non-linear features of the time-series operation data; the triangular basis function is used to obtain the periodic features of the time-series operation data; the radial basis function is used to obtain the non-linear features of the time-series operation data.

[0012] Optionally, the observation space data matrix is fitted to obtain an equivalent fault response model corresponding to the distribution network, including: performing single-step evolution segmentation on the observation space data matrix to obtain an initial data slice set and a data slice set after single-step evolution; performing fitting processing according to the initial data slice set and the data slice set after single-step evolution to obtain an equivalent fault response model corresponding to the distribution network.

[0013] Optionally, based on the equivalent fault response model, a target fault feature vector is extracted, including: performing eigenvalue decomposition on the equivalent fault response model to obtain multiple target eigenvalues; extracting the target fault feature vector according to the multiple target eigenvalues, and the target fault feature vector includes the modal frequency, damping ratio, and modal amplitude corresponding to each target eigenvalue among the multiple target eigenvalues.

[0014] Optionally, based on the target fault feature vector, the target fault type corresponding to the distribution network is identified, including: matching the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network, and the fault feature library is constructed based on the historical time-series operation data and simulation results when short-circuit faults occur in the distribution network, and the fault feature library includes the corresponding relationship between the fault feature vectors and the fault types.

[0015] Optionally, matching the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network includes: using a preset similarity acquisition method to obtain the similarity between the target fault feature vector and the fault feature vectors in the fault feature library, and the preset similarity acquisition method includes at least one of obtaining the Euclidean distance, obtaining the weighted Euclidean distance, or obtaining the cosine similarity; identifying the target fault type corresponding to the distribution network as the fault type of the fault feature vector corresponding to the highest similarity.

[0016] Optionally, the time-series operation data is subjected to dimensionality elevation processing through an observation function to obtain an observation space data matrix, including: preprocessing the time-series operation data to obtain preprocessed time-series operation data, where the preprocessing includes at least one of normalization processing, noise reduction processing, and outlier removal processing; performing dimensionality elevation processing on the preprocessed time-series operation data through the observation function to obtain the observation space data matrix.

[0017] In a second aspect, the present application provides a device for identifying types of short-circuit faults in a distribution network, including:

[0018] An acquisition module, configured to acquire time-series operation data of a bus node corresponding to a faulty line in response to a short-circuit fault occurring in the distribution network;

[0019] A dimensionality elevation processing module, configured to perform dimensionality elevation processing on the time-series operation data through an observation function to obtain an observation space data matrix;

[0020] A fitting processing module, configured to perform fitting processing on the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network;

[0021] An extraction module, configured to extract a target fault feature vector based on the equivalent fault response model;

[0022] An identification module, configured to identify a target fault type corresponding to the distribution network based on the target fault feature vector.

[0023] Optionally, the observation function includes a polynomial basis function, a Gaussian kernel function, a triangular basis function, and a radial basis function. The dimensionality elevation processing module is specifically configured to: input the time-series operation data into the polynomial basis function, the Gaussian kernel function, the triangular basis function, and the radial basis function respectively to perform dimensionality elevation processing on the time-series operation data to obtain the observation space data matrix; where the polynomial basis function is used to obtain the coupling relationship and high-order features between the time-series operation data; the Gaussian kernel function is used to obtain the non-linear features of the time-series operation data; the triangular basis function is used to obtain the periodic features of the time-series operation data; and the radial basis function is used to obtain the non-linear features of the time-series operation data.

[0024] Optionally, the fitting processing module is specifically configured to: perform single-step evolution segmentation on the observation space data matrix to obtain an initial data slice set and a data slice set after single-step evolution; perform fitting processing according to the initial data slice set and the data slice set after single-step evolution to obtain an equivalent fault response model corresponding to the distribution network.

[0025] Optionally, the extraction module is specifically configured to: perform eigenvalue decomposition on the equivalent fault response model to obtain a plurality of target eigenvalues; extract the target fault feature vector according to the plurality of target eigenvalues, where the target fault feature vector includes the modal frequency, damping ratio, and modal amplitude corresponding to each target eigenvalue among the plurality of target eigenvalues.

[0026] Optionally, the recognition module is specifically configured to: match the target fault feature vector with the fault feature vectors in the fault feature library, and identify the target fault type corresponding to the distribution network. The fault feature library is constructed based on the historical time-series operation data and simulation results when short-circuit faults occur in the distribution network, and the fault feature library includes the corresponding relationship between the fault feature vectors and the fault types.

[0027] Optionally, when the recognition module is used to match the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network, it is specifically configured to: obtain the similarity between the target fault feature vector and the fault feature vectors in the fault feature library by using a preset similarity acquisition method. The preset similarity acquisition method includes at least one of obtaining the Euclidean distance, obtaining the weighted Euclidean distance, or obtaining the cosine similarity; identify the target fault type corresponding to the distribution network as the fault type of the fault feature vector corresponding to the highest similarity.

[0028] Optionally, the dimension elevation processing module is specifically configured to: preprocess the time-series operation data to obtain the preprocessed time-series operation data. The preprocessing includes at least one of normalization processing, noise reduction processing, and outlier removal processing; perform dimension elevation processing on the preprocessed time-series operation data through an observation function to obtain an observation space data matrix.

[0029] In a third aspect, the present application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0030] The memory stores computer-executable instructions;

[0031] The processor executes the computer-executable instructions stored in the memory to implement the method for identifying the short-circuit fault type of the distribution network as described in the first aspect of the present application.

[0032] In a fourth aspect, the present application provides a computer-readable storage medium, in which computer program instructions are stored. When the computer program instructions are executed, the method for identifying the short-circuit fault type of the distribution network as described in the first aspect of the present application is implemented.

[0033] In a fifth aspect, the present application provides a computer program product, including a computer program. When the computer program is executed, the method for identifying the short-circuit fault type of the distribution network as described in the first aspect of the present application is implemented.

[0034] A method, device, equipment and medium for identifying types of short - circuit faults in a distribution network provided by this application. In response to a short - circuit fault occurring in the distribution network, the time - series operation data of the bus node corresponding to the faulty line is acquired; the time - series operation data is processed for dimensionality increase through an observation function to obtain an observation - space data matrix, realizing the mapping of the dynamic behavior of a non - linear system to a high - dimensional linear space, so as to better capture complex non - linear dynamic characteristics and obtain richer dynamic response characteristic information of the short - circuit fault in the distribution network in the observation space; the observation - space data matrix is fitted to obtain an equivalent fault response model corresponding to the distribution network. Based on the equivalent fault response model, a target fault feature vector is extracted, realizing the dynamic decomposition and identification of the short - circuit fault in the distribution network in the observation space; furthermore, based on the target fault feature vector, the target fault type corresponding to the distribution network is identified. The amount of calculation is small, and it can more accurately and efficiently identify the types of short - circuit faults in the distribution network, with strong adaptability and real - time performance, and can better meet the requirements of real - time monitoring and rapid response of the distribution network, thus providing effective decision - making support for the operation and management of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The accompanying drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with this application and used together with the specification to explain the principles of this application.

[0036] Figure 1 It is a flowchart of a method for identifying types of short - circuit faults in a distribution network provided by an embodiment of this application;

[0037] Figure 2 It is a flowchart of a method for identifying types of short - circuit faults in a distribution network provided by another embodiment of this application;

[0038] Figure 3 It is a schematic structural diagram of a device for identifying types of short - circuit faults in a distribution network provided by an embodiment of this application;

[0039] Figure 4 It is a schematic structural diagram of an electronic device provided by an embodiment of this application.

[0040] Through the above - mentioned accompanying drawings, the clear embodiments of this application have been shown, and there will be more detailed descriptions hereinafter. These drawings and text descriptions are not intended to limit the scope of the concept of this application in any way, but to explain the concept of this application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0041] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the protection scope of this application.

[0042] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards, and corresponding operation entrances are provided for users to choose to authorize or reject.

[0043] As an infrastructure, the security and stability of the operation of the power system are of crucial importance. During the operation of the power system, faults (such as short circuits, line disconnections, or equipment failures) are inevitable. These faults may lead to system frequency oscillations, voltage instability, or even large-scale power outages. Therefore, rapid detection and accurate analysis of faults are of great significance. Among them, the distribution network is an indispensable part of the power system, responsible for the final distribution and delivery of electricity to ensure that users obtain reliable and stable power supply. When a short-circuit fault occurs in the distribution network, it is of great research significance to quickly and accurately identify the type of short-circuit fault.

[0044] Traditional methods for identifying the types of short-circuit faults in the distribution network usually rely on physical modeling and numerical simulation. For example, time-domain simulation methods, frequency-domain analysis methods, and state-space modeling methods, etc. The above methods rely on the mathematical model of the power system and predict the response behavior of the power system by solving dynamic equations. However, with the continuous improvement of the complexity of the power system, such as the access of new energy, the popularization of distributed generation, and the large-scale application of power electronic devices, the above methods have the following problems: poor adaptability to the non-linear characteristics, complex grid structure, or weak grid of the power system; requiring an accurate grid model and relying on rules and set thresholds; being difficult to cope with the interference caused by high-order harmonics, asymmetry, and dynamic changes.

[0045] Currently, when a short-circuit fault occurs in the distribution network, a data-driven pattern recognition method is usually adopted to identify the type of short-circuit fault. Among them, machine learning or statistical methods are used to extract data features from the operation data of the distribution network for identifying the type of short-circuit fault. The above method performs well when the amount of data is large and the annotation is rich, but it lacks theoretical interpretability and is difficult to be extended to unknown scenarios. Therefore, when identifying the type of short-circuit fault through the above method, there is a problem that the recognition result is not accurate enough.

[0046] In addition, in the related art, a numerical simulation method based on physical modeling can also be used to identify the short-circuit fault types of a distribution network. Among them, relying on a detailed power system dynamics model, the time-domain responses of fault scenarios such as short circuits and oscillations are numerically solved. This method can obtain relatively accurate analysis results, but due to the complexity of non-linearity and high dimensions, the computational cost is relatively high. Alternatively, a method based on hybrid modeling can also be used to identify the short-circuit fault types of a distribution network. Among them, it is necessary to combine a physical model and a data-driven model. For example, the simulation results are used as the input of a machine learning model, or the efficiency of the data-driven method is improved through model simplification. However, the generalization of this method is limited, and it depends on the quality of the physical model.

[0047] In summary, the current methods for identifying short-circuit fault types in distribution networks are insufficient in dealing with the identification of short-circuit fault types after the access of new energy, and it is difficult to quickly and accurately identify different short-circuit fault types and their corresponding dynamic characteristics from the operation data of the distribution network. Moreover, with the increase in the scale and complexity of the distribution network, the computational efficiency of the current methods is difficult to meet the requirements of real-time monitoring and rapid response.

[0048] Based on the above problems, the present application provides a method, device, equipment and medium for identifying short-circuit fault types in a distribution network. When a short-circuit fault occurs in the distribution network, the time-series operation data of the bus node corresponding to the fault line is acquired, and the time-series operation data is dimensionally elevated through an observation function to obtain a higher-dimensional observation space data matrix, so as to map the dynamic behavior of the non-linear system to a high-dimensional linear space, thereby better capturing complex non-linear dynamic characteristics; an equivalent fault response model corresponding to the distribution network is constructed based on the observation space data matrix, and then based on the equivalent fault response model, a target fault feature vector is extracted, and in combination with the extracted target fault feature vector, the target fault type corresponding to the distribution network is identified, so as to more accurately and efficiently identify the short-circuit fault types of the distribution network by combining different observation functions with the adaptive extended dynamic mode (eigenvalue) decomposition technology, with strong adaptability and real-time performance, and can better meet the requirements of real-time monitoring and rapid response of the distribution network, so as to provide effective decision-making support for the operation and management of the distribution network.

[0049] It should be noted that the method for identifying short-circuit fault types in the distribution network provided in the embodiment of the present application can be applied in a server, and the server can be an independent server, or a service cluster, etc.

[0050] Figure 1 This is a flowchart of the method for identifying short-circuit fault types in a distribution network provided in an embodiment of the present application. As Figure 1 shown, the method of the embodiment of the present application includes:

[0051] S101. In response to a short - circuit fault occurring in the distribution network, obtain the time - series operation data of the bus node corresponding to the faulty line.

[0052] In the embodiments of the present application, the bus node is a key component in the power system, which is used to collect and distribute electric power. The distribution network contains multiple bus nodes. Exemplarily, when a short - circuit fault occurs in the distribution network, the time - series operation data of the bus node corresponding to the faulty line can be obtained (collected). Specifically, for example, the time - series operation data of the bus node corresponding to the faulty line within a preset time period before and after the short - circuit fault in the distribution network can be obtained. Among them, the time - series operation data is, for example, time - series data of different dimensions (types) such as voltage amplitude, phase angle, frequency corresponding to different moments, and negative - sequence voltage components and zero - sequence voltage components reflecting the asymmetric operation state of the distribution network. The above - mentioned multi - dimensional time - series operation data can provide comprehensive distribution network operation state information.

[0053] Optionally, considering that the obtained time - series operation data is usually affected by noise, outliers, and dimensional differences, directly using it may lead to a decrease in analysis accuracy. Therefore, the obtained time - series operation data can be pre - processed to obtain pre - processed time - series operation data. The pre - processing includes at least one of normalization processing, noise reduction processing, and outlier removal processing. Thus, based on the pre - processed time - series operation data, the subsequent steps can be continued, which can effectively improve the comparability of feature data under the same dimension. For how to specifically pre - process the time - series operation data, reference can be made to the subsequent embodiments and will not be elaborated here.

[0054] S102. Perform dimensionality - elevation processing on the time - series operation data through an observation function to obtain an observation - space data matrix.

[0055] It can be understood that by performing dimensionality - elevation processing on the time - series operation data through the constructed observation function, the dynamic behavior of the non - linear system can be mapped to a high - dimensional linear space, so as to better capture complex non - linear dynamic features. The observation function includes, for example, polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions. Different observation functions can help match specific fault types and dynamic features. In this step, the time - series operation data can be subjected to dimensionality - elevation processing through the observation function to obtain a higher - dimensional observation - space data matrix, that is, more abundant distribution network short - circuit fault dynamic response feature information is obtained in the observation space. For how to specifically obtain the observation - space data matrix, reference can be made to the subsequent embodiments and will not be elaborated here.

[0056] S103. Perform fitting processing on the observation - space data matrix to obtain an equivalent fault response model corresponding to the distribution network.

[0057] Exemplarily, according to the observation space data matrix, an observation space single-step evolution slice data set can be constructed, and then fitting processing can be performed according to the observation space single-step evolution slice data set to obtain an equivalent fault response model corresponding to the distribution network. For how to obtain the equivalent fault response model specifically, reference can be made to the subsequent embodiments, which will not be elaborated here.

[0058] S104. Extract the target fault feature vector based on the equivalent fault response model.

[0059] In this step, eigenvalue decomposition can be performed on the equivalent fault response model to extract the target fault feature vector. For how to extract the target fault feature vector based on the equivalent fault response model specifically, reference can be made to the subsequent embodiments, which will not be elaborated here.

[0060] S105. Identify the target fault type corresponding to the distribution network based on the target fault feature vector.

[0061] Exemplarily, after the target fault feature is extracted, the target fault feature vector can be matched with the fault feature vectors in the pre-constructed fault feature library to identify the target fault type corresponding to the distribution network. For example, the short-circuit fault types include single-phase high-resistance grounding fault, interphase short-circuit fault, two-phase grounding short-circuit fault, open-circuit short-circuit, and three-phase short-circuit, etc. Then the target fault type is specifically, for example, three-phase short-circuit. After identifying the target fault type corresponding to the distribution network, the target fault type can be output to facilitate corresponding processing in a timely manner according to the target fault type and solve the short-circuit fault problem of the distribution network.

[0062] The method for identifying the short-circuit fault type of the distribution network provided by the embodiments of the present application, by responding to a short-circuit fault occurring in the distribution network, obtains the time-series operation data of the bus node corresponding to the faulty line; performs dimensionality elevation processing on the time-series operation data through an observation function to obtain an observation space data matrix, realizing mapping the dynamic behavior of the nonlinear system to a high-dimensional linear space, so as to better capture complex nonlinear dynamic characteristics and obtain richer dynamic response characteristic information of the short-circuit fault of the distribution network in the observation space; performs fitting processing on the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network, extracts the target fault feature vector based on the equivalent fault response model, realizing dynamic decomposition and identification of the short-circuit fault of the distribution network in the observation space; and then based on the target fault feature vector, identifies the target fault type corresponding to the distribution network, with a small amount of calculation, can more accurately and efficiently identify the short-circuit fault type of the distribution network, has strong adaptability and real-time performance, and better meets the requirements of real-time monitoring and rapid response of the distribution network, so as to provide effective decision-making support for the operation and management of the distribution network.

[0063] Figure 2The flowchart of the method for identifying the types of short - circuit faults in a distribution network provided by another embodiment of this application. On the basis of the above - mentioned embodiment, the embodiment of this application further describes the method for identifying the types of short - circuit faults in a distribution network. As Figure 2 shown, the method of the embodiment of this application may include:

[0064] S201. In response to a short - circuit fault occurring in the distribution network, obtain the time - series operation data of the bus node corresponding to the faulty line.

[0065] For the specific description of this step, reference may be made to the relevant description of S101 in the Figure 1 shown embodiment, which will not be elaborated here.

[0066] Considering the pre - processing of the time - series operation data, and the observation function includes polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions. Therefore, in the embodiment of this application, Figure 1 step S102 in

[0067] may further include the following two steps of S202 and S203:

[0068] Exemplarily, considering that the obtained time - series operation data (which can also be called the original measurement data) is usually affected by noise, outliers, and dimensional differences, directly using it may lead to a decrease in the analysis accuracy. Therefore, the obtained time - series operation data can be pre - processed to obtain the pre - processed time - series operation data. Among them, the pre - processing can, for example, include normalization processing, noise reduction processing, and outlier removal processing. The following will elaborate on different pre - processing methods respectively.

[0069] (1) Normalization processing: By scaling different data features to a unified range, the differences between different dimensions can be eliminated. In the embodiment of this application, for example, the softmax function (a type of normalized exponential function) can be used to perform normalization processing on the time - series operation data, and the corresponding formula is Formula 1 below:

[0070]

[0071] where z j represents the measured value of the original measurement data (such as voltage amplitude, etc.) of a certain dimension (type) at the j - th moment; x j represents the measured value of the original measurement data of a certain dimension (type) after normalization processing at the j - th moment; N represents the total time length corresponding to the original measurement data of a certain dimension (type).

[0072] (2) Noise reduction processing: The noise reduction processing is to eliminate the influence of measurement errors and random interferences on the time-series operation data. In the embodiments of the present application, for example, a Kalman filter can be used for noise reduction processing to gradually estimate the true value by using the state space model of the data. For the specific method of using the Kalman filter for noise reduction processing, reference can be made to the current related technologies. It can be understood that the Kalman filter can dynamically smooth the data in the time series, eliminate the noise and obtain an estimate close to the true value.

[0073] (3) Outlier removal processing: In the embodiments of the present application, for example, the 3σ criterion based on the mean and standard deviation can be adopted. This method uses the statistical characteristics of the data to screen out the outliers. For example, for the time-series operation data x = [x 1 , x 2 ,..., x N after normalization processing for a certain dimension (type), the mean μ and the standard deviation σ can be obtained through the following formula two:

[0074]

[0075] Among them, μ represents the mean of the time-series operation data of a certain dimension (type); σ represents the standard deviation of the time-series operation data of a certain dimension (type).

[0076] According to the 3σ criterion (three times the standard deviation principle), the condition for judging whether a data point is an outlier is: |x i - μ| > 3σ. The data point x i that satisfies this condition is determined as an outlier, and the outlier is removed or replaced with the mean μ or the interpolation of adjacent points.

[0077] It can be understood that through the above normalization processing, noise reduction processing and outlier removal processing, high-quality time-series operation data can be obtained. The characteristics of this time-series operation data can include: the data is uniformly scaled to a reasonable range, eliminating the dimensional difference between features; through the Kalman filter, high-frequency noise interference is eliminated, enhancing the stability of the data; after removing the outliers, the data more truly reflects the operation state of the distribution network, providing a reliable input for the subsequent extended dynamic mode decomposition.

[0078] S203. Respectively input the preprocessed time-series operation data into the polynomial basis function, Gaussian kernel function, triangular basis function, and radial basis function to perform dimensionality elevation processing on the preprocessed time-series operation data, and obtain an observation space data matrix.

[0079] Among them, the polynomial basis function is used to obtain the coupling relationship and high-order features between the time-series operation data; the Gaussian kernel function is used to obtain the non-linear features of the time-series operation data; the triangular basis function is used to obtain the periodic features of the time-series operation data; the radial basis function is used to obtain the non-linear features of the time-series operation data.

[0080] It can be understood that by performing dimensionality elevation processing on the preprocessed time-series operation data through a constructed observation function, mapping the dynamic behavior of the nonlinear system to a high-dimensional linear space can better capture complex nonlinear features. Among them, the selection of the observation function is crucial, and different observation functions can help match specific fault types and dynamic features.

[0081] Exemplarily, it is assumed that the preprocessed time-series operation data is represented in the form of a time series as follows:

[0082]

[0083] where m represents the dimension (type) of the data, for example, the voltage amplitude is one of the dimensions (types); N represents the number of time steps. The different observation functions will be described in detail below.

[0084] (1) Polynomial basis function, which is used to obtain the coupling relationship and high-order features between the preprocessed time-series operation data, that is, it is applicable to describe nonlinear behaviors with polynomial characteristics, such as symmetric voltage fluctuations or other symmetry faults; the polynomial basis function can be expressed as:

[0085] where X is the preprocessed time-series operation data; x 1 , x 2 ,..., x m represent the measurement data of the 1st dimension (type) to the mth dimension (type) in the preprocessed time-series operation data, and the measurement data of each dimension (type) is a row vector; T represents the transpose.

[0086] (2) Gaussian kernel function, which is used to obtain the nonlinear features of the time-series operation data, that is, it is applicable to obtain local dynamic characteristics, such as the dynamic characteristics caused by local line faults or short circuits; the Gaussian kernel function can be expressed as:

[0087]

[0088] where x represents the measurement data of a certain dimension (type) in the preprocessed time-series operation data, which is a row vector; c represents the kernel center corresponding to the measurement data of this dimension (type), usually selected as the clustering center of the data; η is the kernel width parameter, which controls the local characteristics.

[0089] (3) Triangular basis function, which is used to obtain the periodic features of the time-series operation data, that is, it is applicable to capture periodic dynamic behaviors, such as frequency oscillation faults or subsynchronous resonance phenomena; the triangular basis function can be expressed as: Φ(x) = [sin(ω 1 x) T , cos(ω1 x) T ,..., sin(ω k x) T , cos(ω x ) T T , where x represents the measurement data of a certain dimension (type) in the preprocessed time-series operation data, which is a row vector; ω 1 represents the fundamental frequency of the distribution network, such as 50 Hz; ω k represents the kth harmonic frequency of the distribution network, that is, ω k = kω 1 .

[0090] (4) Radial basis function, which is used to obtain the non-linear characteristics of time-series operation data, that is, it is suitable for obtaining asymmetric non-linear dynamics, such as single-phase grounding short circuit, two-phase short circuit, etc.; the radial basis function can be expressed as:

[0091]

[0092] where x represents the measurement data of a certain dimension (type) in the preprocessed time-series operation data, which is a row vector; τ represents the center of the basis function corresponding to the measurement data of this dimension (type), usually a representative point selected from the input data; ψ represents the scale parameter, which is used to determine the width of the basis function response.

[0093] It can be understood that in the selection of the observation function to match the fault type, the polynomial basis function is applicable to global faults, such as distribution network frequency drift or voltage sag faults. The Gaussian kernel function is applicable to the analysis of short-term local faults, such as single-phase short circuit or local bus anomalies. The triangular basis function is dedicated to the analysis of periodic behaviors, such as frequency oscillation or subsynchronous resonance faults. The radial basis function is applicable to dealing with asymmetric dynamic characteristics and is suitable for the change analysis of negative sequence or zero sequence components. Through the above observation functions, a flexible dimension elevation strategy can be realized, and an observation space data matrix of the distribution network fault response can be constructed, which can also be called an observation space data set.

[0094] In this step, the preprocessed time-series operation data can be respectively input into the polynomial basis function, Gaussian kernel function, triangular basis function, and radial basis function to perform dimension elevation processing on the preprocessed time-series operation data, and a higher-dimensional observation space data matrix can be obtained. For example, it is represented by X′, and X′ is an n×N matrix, where n represents the dimension (type) of the data in the observation space data matrix, that is, the number of rows of X′, n is greater than m, and N is the number of time steps.

[0095] In the embodiment of the present application, Figure 1 step S103 in can further include the following two steps of S204 and S205: ​

[0096] S204. Perform a single-step evolution segmentation on the observed space data matrix to obtain an initial data slice set and a data slice set after single-step evolution.

[0097] Exemplarily, a single-step evolution segmentation can be performed on the observed space data matrix. That is, the observed space data from the 1st time step to the (N - 1)th time step constitutes the initial data slice set (for example, represented by Φ(X′)′), and the observed space data from the 2nd time step to the Nth time step constitutes the data slice set after single-step evolution (for example, represented by Φ(X′)″). That is: Φ(X′)′ = Φ(X′)[1:N - 1], Φ(X′)″ = Φ(X′)[2:N].

[0098] S205. Perform a fitting process based on the initial data slice set and the data slice set after single-step evolution to obtain an equivalent fault response model corresponding to the distribution network.

[0099] Exemplarily, based on the single-step evolution relationship of the observed space data matrix, a least squares method can be used for fitting based on the initial data slice set and the data slice set after single-step evolution to obtain an equivalent fault response model K corresponding to the distribution network. Among them, the single-step evolution relationship satisfies: Φ(X′)″≈KΦ(X′)′. Therefore, the obtained equivalent fault response model K corresponding to the distribution network is: K = Φ(X′)″Φ(X′)′ + , where Φ(X′)′ + is the pseudo-inverse of Φ(X′)′.

[0100] In the embodiments of the present application, Figure 1 Step S104 can further include the following two steps of S206 and S207:

[0101] S206. Perform eigenvalue decomposition on the equivalent fault response model to obtain multiple target eigenvalues.

[0102] Exemplarily, the equivalent fault response model can be eigenvalue decomposed through the following formula three to obtain multiple target eigenvalues, for example, represented by λ i , where i is an integer greater than or equal to 1 and less than or equal to n. At the same time, multiple eigenvectors can also be obtained, for example, represented by φ i .

[0103] Kφ i = λ i φ i Formula Three

[0104] S207. Extract a target fault eigenvector according to the multiple target eigenvalues. The target fault eigenvector includes the modal frequency, damping ratio, and modal amplitude corresponding to each target eigenvalue among the multiple target eigenvalues.

[0105] Exemplarily, the modal frequency corresponding to each target eigenvalue can be obtained through the following Formula 4 (for example, represented by ω i ):

[0106]

[0107] The damping ratio corresponding to each target eigenvalue can be obtained through the following Formula 5 (for example, represented by ζ i ):

[0108]

[0109] The modal amplitude (for example, represented by a i ) is the amplitude of the target eigenvalue, thereby obtaining the target fault feature vector.

[0110] In the embodiments of the present application, Figure 1 Step S105 in can further include the following Step S208:

[0111] S208. Match the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network.

[0112] Among them, the fault feature library is constructed based on the historical time-series operation data and simulation results when a short-circuit fault occurs in the distribution network, and the fault feature library includes the corresponding relationship between the fault feature vector and the fault type.

[0113] Exemplarily, based on the example of Step S206, if the number of target eigenvalues is n, the feature vector corresponding to each target eigenvalue can be expressed as: v dynamic，i =[ω i , ζ i , a i , where i is an integer greater than or equal to 1 and less than or equal to n. Then, the target fault feature vector composed of the feature vectors corresponding to all n target eigenvalues can be expressed as:

[0114] v dynamic =[v dynamic，1 , v dynamic，2 ,..., v dynamic，n T =[ω 1 , ζ 1 , a 1 , ω 2 , ζ 2 , a 2 ,..., ω n , ζ n , a n T

[0115] ​​A fault feature library can be constructed based on historical sequential operation data and simulation results when a short - circuit fault occurs in the distribution network. The fault feature library includes the correspondence between fault feature vectors and fault types, that is, each fault feature vector corresponds to a fault type. For the j - th fault type, each corresponding fault feature vector can be expressed as: v fault，ji =[ω fault，ji , ζ fault，ji , ζ faul，ji , a fault，ji . Then, the feature vector composed of all n fault feature vectors of the j - th fault type can be expressed as:

[0116] v fault，j =[v fault，ji , v fault，j2 , …, v fault，jn

[0117] =[ω fault，j1 , ζ fault，j1 , a fault，j1 , ω fault，j2 , ζ fault，j2 , a fault，j2 ,... ω fault，jn , ζ fault，jn , a fault，jn T

[0118] where j represents the j - th fault type.

[0119] After extracting the target fault feature vector, the target fault feature vector can be matched with the fault feature vectors in the pre - constructed fault feature library to identify the target fault type of the distribution network. For example, short - circuit fault types include single - phase high - resistance grounding fault, inter - phase short - circuit fault, two - phase grounding short - circuit fault, open - wire short - circuit, and three - phase short - circuit, etc. Then the target fault type could be, for example, three - phase short - circuit.

[0120] Furthermore, optionally, matching the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type of the distribution network may include: obtaining the similarity between the target fault feature vector and the fault feature vectors in the fault feature library by using a preset similarity - obtaining method, where the preset similarity - obtaining method includes at least one of obtaining Euclidean distance, obtaining weighted Euclidean distance, or obtaining cosine similarity; identifying the target fault type of the distribution network as the fault type corresponding to the fault feature vector with the highest similarity.

[0121] Exemplarily, the Euclidean distance can be used to measure the difference between the dynamic modal features and various fault features. Specifically, the Euclidean distance between the target fault feature vector and the fault feature vectors in the fault feature library can be obtained through the following formula (6): ​​

[0122]

[0123] Alternatively, the weighted Euclidean distance between the target fault feature vector and the fault feature vectors in the fault feature library can be obtained through the following formula seven, where weights γ can be assigned to the feature vectors corresponding to different target feature values. i :

[0124]

[0125] where γ i can be dynamically adjusted according to the importance of the target feature value. For example, the weight of frequency ω is higher in frequency oscillation faults.

[0126] Alternatively, the cosine similarity between the target fault feature vector and the fault feature vectors in the fault feature library can be obtained through the following formula eight, that is, the direction consistency of the fault feature vectors is measured by the cosine similarity:

[0127]

[0128] After obtaining the similarity between the target fault feature vector and the fault feature vectors in the fault feature library, select the one with the highest similarity v fault，j , for example: where the fault feature vector corresponding to j * is the matching result, so that the target fault type corresponding to the distribution network can be identified as the fault type to which the fault feature vector with the highest similarity belongs.

[0129] Combined with the matching result of the fault feature vector, the following common fault types can be identified:

[0130] (1) Single-phase grounding short-circuit fault, with characteristics: modal frequency ω i within a specific interval (such as 0.2 Hz - 2 Hz); modal amplitude a i is larger; damping ζ i is close to zero or negative; the corresponding identification method is: if the modal frequency and damping ratio in the matching result simultaneously satisfy: ω fault ∈ [ω lower , ω upper , ζ fault ≤ 0, where ω lower represents the minimum value, and ω upper represents the maximum value, then it can be determined that there is a single-phase grounding short-circuit fault in the distribution network.

[0131] (2) Three-phase short circuit, with characteristics: modal amplitude a idecreases and the voltage component corresponding to the bus bar drops significantly; the corresponding identification method is: if the modal amplitude in the matching result satisfies: a fault < a threshold , where a threshold represents the threshold of the modal amplitude, then it can be determined that there is a three-phase short circuit fault in the distribution network.

[0132] (3) Other asymmetrical short circuit faults, characterized by: the modal frequency ω i in the high frequency range (such as above 50 Hz); the modal amplitude and modal frequency fluctuate violently in a short time; the damping ratio ζ i decreases; the corresponding identification method is: if the modal frequency and damping ratio in the matching result simultaneously satisfy: ω fault > ω threshold , ζ fault < ζ threshold , where ω threshold represents the threshold of the modal frequency, and ζ threshold represents the threshold of the damping ratio, then it can be determined that there is an asymmetrical short circuit fault in the distribution network.

[0133] It can be understood that in the embodiments of the present application, by matching the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network, it has the advantages of small calculation amount, fast calculation efficiency, higher accuracy and easy expansion, can quickly identify the target fault type, improve the accuracy of identifying the short circuit fault type of the distribution network, and provide important support for the operation monitoring and fault emergency handling of the distribution network.

[0134] The method for identifying the type of short - circuit fault in a distribution network provided by an embodiment of the present application includes: in response to a short - circuit fault occurring in the distribution network, obtaining the time - series operation data of the bus node corresponding to the faulty line; pre - processing the time - series operation data to obtain the pre - processed time - series operation data, where the pre - processing includes at least one of normalization processing, noise reduction processing, and outlier rejection processing, which can effectively improve the comparability of feature data under the same dimension; respectively inputting the pre - processed time - series operation data into polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions to perform dimensionality - raising processing on the pre - processed time - series operation data, obtaining an observation space data matrix, and realizing mapping the dynamic behavior of a non - linear system to a high - dimensional linear space, so as to better capture complex non - linear dynamic features and obtain richer dynamic response feature information of the short - circuit fault in the distribution network in the observation space; performing single - step evolution segmentation on the observation space data matrix to obtain an initial data slice set and a data slice set after single - step evolution, performing fitting processing according to the initial data slice set and the data slice set after single - step evolution to obtain an equivalent fault response model corresponding to the distribution network; performing eigenvalue decomposition on the equivalent fault response model to obtain multiple target eigenvalues, and extracting a target fault feature vector according to the multiple target eigenvalues. The target fault feature vector includes the modal frequency, damping ratio, and modal amplitude corresponding to each target eigenvalue among the multiple target eigenvalues, realizing dynamic decomposition and identification of the short - circuit fault of the distribution network in the observation space; matching the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network. The amount of calculation is small and the calculation efficiency is relatively fast, so that the type of short - circuit fault in the distribution network can be identified more accurately and efficiently, with strong adaptability and real - time performance, and can provide effective decision - making support for the operation and management of the distribution network, especially can provide important support for the operation monitoring and fault emergency handling of the distribution network.

[0135] The following is an embodiment of the device of the present application, which can be used to execute the method embodiment of the present application. For the details not disclosed in the embodiment of the device of the present application, please refer to the method embodiment of the present application.

[0136] Figure 3 It is a schematic structural diagram of an identification device for the type of short - circuit fault in a distribution network provided by an embodiment of the present application. As Figure 3 shown, the identification device 300 for the type of short - circuit fault in a distribution network according to an embodiment of the present application includes: an acquisition module 301, a dimensionality - raising processing module 302, a fitting processing module 303, an extraction module 304, and an identification module 305. Among them:

[0137] The acquisition module 301 is configured to obtain the time - series operation data of the bus node corresponding to the faulty line in response to a short - circuit fault occurring in the distribution network.

[0138] The dimensionality - raising processing module 302 is configured to perform dimensionality - raising processing on the time - series operation data through an observation function to obtain an observation space data matrix.

[0139] The fitting processing module 303 is configured to perform fitting processing on the observed space data matrix to obtain an equivalent fault response model corresponding to the distribution network.

[0140] The extraction module 304 is configured to extract a target fault feature vector based on the equivalent fault response model.

[0141] The identification module 305 is configured to identify a target fault type corresponding to the distribution network based on the target fault feature vector.

[0142] In some embodiments, the observation function includes polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions. The dimension elevation processing module 302 may specifically be configured to: input the time-series operation data into the polynomial basis functions, Gaussian kernel functions, triangular basis functions, and radial basis functions respectively to perform dimension elevation processing on the time-series operation data to obtain an observed space data matrix; wherein, the polynomial basis function is used to obtain the coupling relationship and high-order features between the time-series operation data; the Gaussian kernel function is used to obtain the non-linear features of the time-series operation data; the triangular basis function is used to obtain the periodic features of the time-series operation data; and the radial basis function is used to obtain the non-linear features of the time-series operation data.

[0143] In some embodiments, the fitting processing module 303 may specifically be configured to: perform single-step evolution segmentation on the observed space data matrix to obtain an initial data slice set and a data slice set after single-step evolution; perform fitting processing according to the initial data slice set and the data slice set after single-step evolution to obtain an equivalent fault response model corresponding to the distribution network.

[0144] In some embodiments, the extraction module 304 may specifically be configured to: perform eigenvalue decomposition on the equivalent fault response model to obtain a plurality of target eigenvalues; extract a target fault feature vector according to the plurality of target eigenvalues, and the target fault feature vector includes the modal frequency, damping ratio, and modal amplitude corresponding to each target eigenvalue among the plurality of target eigenvalues.

[0145] In some embodiments, the identification module 305 may specifically be configured to: match the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network. The fault feature library is constructed based on the historical time-series operation data and simulation results when short-circuit faults occur in the distribution network, and the fault feature library includes the corresponding relationship between the fault feature vectors and the fault types.

[0146] Optionally, when the recognition module 305 is used to match the target fault feature vector with the fault feature vectors in the fault feature library to identify the target fault type corresponding to the distribution network, it can specifically be used to: obtain the similarity between the target fault feature vector and the fault feature vectors in the fault feature library by using a preset similarity acquisition method, where the preset similarity acquisition method includes at least one of obtaining the Euclidean distance, obtaining the weighted Euclidean distance, or obtaining the cosine similarity; identify the target fault type corresponding to the distribution network as the fault type of the fault feature vector corresponding to the highest similarity.

[0147] Optionally, the dimensionality increase processing module 302 can specifically be used to: preprocess the time-series operation data to obtain the preprocessed time-series operation data, where the preprocessing includes at least one of normalization processing, noise reduction processing, and outlier removal processing; perform dimensionality increase processing on the preprocessed time-series operation data through an observation function to obtain an observation space data matrix.

[0148] The device in the embodiments of the present application can be used to execute the technical solutions in any of the above-described method embodiments, and its implementation principles and technical effects are similar, which will not be elaborated here.

[0149] Figure 4 This is a schematic structural diagram of an electronic device provided in an embodiment of the present application. As Figure 4 shown, the electronic device 400 may include: at least one processor 401 and a memory 402.

[0150] The memory 402 is used to store a program. Specifically, the program may include program code, and the program code includes computer execution instructions.

[0151] The memory 402 may include a high-speed random access memory (Random Access Memory, RAM), and may also include a non-volatile memory, such as at least one disk memory.

[0152] The processor 401 is used to execute the computer execution instructions stored in the memory 402 to implement the method for identifying the short-circuit fault type of the distribution network described in the foregoing method embodiments. Among them, the processor 401 may be a central processing unit (Central Processing Unit, CPU), or an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits configured to implement the embodiments of the present application. Specifically, when implementing the method for identifying the short-circuit fault type of the distribution network described in the foregoing method embodiments, the electronic device may be an electronic device with processing functions such as a server.

[0153] Optionally, the electronic device 400 may further include a communication interface 403. In a specific implementation, if the communication interface 403, the memory 402, and the processor 401 are implemented independently, the communication interface 403, the memory 402, and the processor 401 may be interconnected through a bus and communicate with each other. The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc., but it does not mean that there is only one bus or one type of bus.

[0154] Optionally, in a specific implementation, if the communication interface 403, the memory 402, and the processor 401 are integrated on a single chip, the communication interface 403, the memory 402, and the processor 401 may communicate through an internal interface.

[0155] This application also provides a computer-readable storage medium storing computer program instructions, and when the processor executes the computer program instructions, the solution of the above method for identifying the type of short-circuit fault in a distribution network is implemented.

[0156] This application also provides a computer program product including a computer program, and when the computer program is executed, the solution of the above method for identifying the type of short-circuit fault in a distribution network is implemented.

[0157] The above computer-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a Static Random Access Memory (SRAM), an Electrically Erasable Programmable Read Only Memory (EEPROM), an Erasable Programmable Read Only Memory (EPROM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a magnetic memory, a flash memory, a magnetic disk, or an optical disc. The readable storage medium may be any available medium accessible by a general-purpose or special-purpose computer.

[0158] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit. Of course, the processor and the readable storage medium can also exist as discrete components in a device for identifying the types of short-circuit faults in a distribution network.

[0159] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for identifying the type of short-circuit fault in a distribution network, characterized in that: include: In response to a short circuit fault in the distribution network, obtaining time-series operation data of a bus node corresponding to the fault line; Performing dimension-upgrading processing on the time series operation data through an observation function to obtain an observation space data matrix; Performing fitting processing on the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network; Based on the equivalent fault response model, extracting a target fault feature vector; Based on the target fault feature vector, a target fault type corresponding to the distribution network is identified.

2. The identification method according to claim 1, characterized in that: The observation function includes a polynomial basis function, a Gaussian kernel function, a triangular basis function and a radial basis function. The time series operation data is subjected to dimensionality-increasing processing by the observation function to obtain an observation space data matrix, including: Inputting the time series operation data into the polynomial basis function, the Gaussian kernel function, the triangular basis function and the radial basis function respectively to perform dimension-raising processing on the time series operation data to obtain an observation space data matrix; Among them, the polynomial basis function is used to obtain the coupling relationship and high-order characteristics between the timing operation data; the Gaussian kernel function is used to obtain the nonlinear characteristics of the timing operation data; the triangular basis function is used to obtain the periodic characteristics of the timing operation data; and the radial basis function is used to obtain the nonlinear characteristics of the timing operation data.

3. The identification method according to claim 2, characterized in that: The fitting process of the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network includes: Performing single-step evolution segmentation on the observation space data matrix to obtain an initial data slice set and a single-step evolved data slice set; Fitting processing is performed according to the initial data slice set and the single-step evolved data slice set to obtain an equivalent fault response model corresponding to the distribution network.

4. The identification method according to claim 3, characterized in that: The step of extracting a target fault feature vector based on the equivalent fault response model includes: Performing eigenvalue decomposition on the equivalent fault response model to obtain multiple target eigenvalues; The target fault feature vector is extracted according to the multiple target feature values, and the target fault feature vector includes a modal frequency, a damping ratio, and a modal amplitude corresponding to each target feature value in the multiple target feature values.

5. The identification method according to any one of claims 1 to 4, characterized in that: The identifying, based on the target fault feature vector, a target fault type corresponding to the distribution network includes: The target fault feature vector is matched with the fault feature vector in the fault feature library to identify the target fault type corresponding to the distribution network. The fault feature library is constructed based on the historical timing operation data and simulation results when a short circuit fault occurs in the distribution network. The fault feature library includes the correspondence between the fault feature vector and the fault type.

6. The identification method according to claim 5, characterized in that: The matching of the target fault feature vector with the fault feature vector in the fault feature library to identify the target fault type corresponding to the distribution network includes: A preset similarity acquisition method is used to acquire the similarity between the target fault feature vector and the fault feature vector in the fault feature library, wherein the preset similarity acquisition method includes at least one of acquiring Euclidean distance, acquiring weighted Euclidean distance, or acquiring cosine similarity; The target fault type corresponding to the distribution network is identified as the fault type of the fault feature vector corresponding to the highest similarity.

7. The identification method according to any one of claims 1 to 4, characterized in that: The step of performing dimension-upgrading processing on the time series operation data by using an observation function to obtain an observation space data matrix includes: Preprocessing the time series operation data to obtain preprocessed time series operation data, wherein the preprocessing includes at least one of normalization processing, noise reduction processing and outlier elimination processing; The pre-processed time series operation data is subjected to dimensionality increase processing through an observation function to obtain an observation space data matrix.

8. A device for identifying short-circuit fault types in a distribution network, characterized in that: include: An acquisition module, used for acquiring the time series operation data of the bus node corresponding to the fault line in response to a short circuit fault in the distribution network; A dimension-increasing processing module is used to perform dimension-increasing processing on the time series operation data through an observation function to obtain an observation space data matrix; A fitting processing module, used for performing fitting processing on the observation space data matrix to obtain an equivalent fault response model corresponding to the distribution network; An extraction module, used for extracting a target fault feature vector based on the equivalent fault response model; An identification module is used to identify a target fault type corresponding to the distribution network based on the target fault feature vector.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method for identifying the type of short-circuit fault in a distribution network according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed, the method for identifying the type of short-circuit fault in the distribution network according to any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed, the method for identifying the type of short-circuit fault in a power distribution network according to any one of claims 1 to 7 is implemented.

Images