Intelligent identification method for topology abnormity of power distribution network

Abstract

The invention relates to a power distribution network topology abnormity intelligent identification method comprising the following steps: S1, constructing a topology structure of a power distribution network, and forming a connection relation graph between nodes; s2, constructing an association relationship between a measurement object file and measurement data of each node of the power distribution network; s3, on the basis of the topological structure of the power distribution network, traversing the graph through a depth-first search method, obtaining the structure and characteristics of the power distribution network in combination with a centrality analysis and community discovery method, and identifying important nodes and community structures; s4, carrying out detection based on the anomaly detection model, carrying out anomaly detection on the topological data of the power distribution network, and obtaining an abnormal node; s5, based on an exception type identification model, acquiring an exception type of the exception node; and S6, in combination with the important nodes and the community structure, evaluating the severity level of the abnormity according to the influence degree of the abnormal nodes and the positions of the abnormal nodes in the network, obtaining the severity level of the abnormity, and performing visualization. The operation safety and the management efficiency of the power distribution network can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of distribution network anomaly detection, and in particular to a method for intelligently identifying distribution network topology anomalies. Background Art

[0002] With the development of distribution networks and their widespread application in the power Internet of Things, in addition to completing basic power distribution tasks, distribution networks also need to perform tasks such as information sharing and equipment monitoring. The correctness of the distribution network topology is the basic prerequisite for realizing these functions, so the verification of the distribution network topology is extremely important. In the prior art, the topology of the distribution network may be abnormal, and there may be topological errors, that is, there are errors in the connection relationship or topological structure between nodes in the network, which may cause failures and problems during the operation of the power grid, affecting the reliability of power supply. There are also loop problems that may cause abnormal current distribution and voltage instability, increasing the risk of power grid operation. Loop problems may also lead to increased grid losses and reduced system efficiency. Summary of the invention

[0003] In order to solve the above problems, the purpose of the present invention is to provide a method for intelligently identifying distribution network topology anomalies, which can effectively improve the operation safety and management efficiency of the distribution network.

[0004] To achieve the above object, the present invention adopts the following technical solutions: A method for intelligently identifying abnormal topology of a distribution network comprises the following steps: S1: Abstract power plants, substations, distribution lines, distribution transformers, and users as nodes, which are represented by unique identifiers. According to the connection relationship between devices, branches are defined to represent the power transmission path between devices. According to the connection relationship between nodes and branches, the topological structure of the distribution network is constructed to form a connection relationship diagram between nodes. S2: Construct the association relationship between the measurement object files and measurement data of each node in the distribution network; S3: Based on the topological structure of the distribution network, the graph is traversed by the depth-first search method, and the structure and characteristics of the distribution network are obtained by combining centrality analysis and community discovery methods, and important nodes and community structures are identified; S4: Collect the features of real-time distribution network topology data, and perform detection based on the anomaly detection model, perform anomaly detection on the distribution network topology data, and obtain abnormal nodes; S5: Obtain the measurement object file and measurement data of the abnormal node, and obtain the abnormal type of the abnormal node based on the abnormal type recognition model; S6: Based on the abnormal nodes and their abnormal types, combined with important nodes and community structures, the severity of the abnormality is evaluated according to the impact of the abnormal nodes and their position in the network, the severity of the abnormality is obtained, and it is visualized.

[0005] Furthermore, S2 is specifically as follows: the measurement object file includes the device information, location, and parameter settings of the node, and is associated with the measurement data through a unique identifier; each measurement object file corresponds to real-time or historically collected measurement data, which is used to describe the device status and power system operation status; the physical connection relationship between the devices in the measurement object file is described through the topological connection relationship; the measurement data is associated through the topological connection relationship in the measurement object file.

[0006] Furthermore, S3 is specifically: Starting from a starting node, recursively visit its neighbor nodes and mark the visited nodes until all nodes have been visited; Calculate the degree centrality, betweenness centrality and closeness centrality of nodes and identify important nodes; ; ;

[0007] in, is the degree centrality of node v; is the degree of node v, that is, the number of edges directly connecting node v to other nodes; N is the total number of nodes; represents betweenness centrality; Represents the number of shortest paths from node s to node t; represents the number of shortest paths passing through node v; represents the closeness centrality, Represents the shortest path length from node v to node u; Using modularity algorithm for community discovery to identify community structures in distribution networks; ; Among them, Q is the modularity, represents the connection weight between node u and node v, , They represent the degrees of nodes u and v respectively, m represents the total number of edges, is an indicator function, which is 1 if node u and node v belong to the same community, otherwise it is 0.

[0008] Furthermore, the anomaly detection model is built based on the GraphSAGE model, as follows: Build a GraphSAGE model to learn node representations and capture the relationships between nodes. Update the node representation by sampling neighbor node features and aggregating them. In the GraphSAGE model, the representation update of node v is calculated as follows: ; ; in, is the aggregation result of the representation of the neighbor nodes of node v at the lth layer; is the representation of node v at layer l; is the set of neighbor nodes of node v; is the neighbor node feature aggregation function, specifically ; is the weight matrix of the lth layer; is the activation function; CONCAT is the concatenation function, which means that the node v is The representation of the layer is concatenated with the representation of the neighbor nodes of node v at layer l; In the feature aggregation stage, a multi-scale feature aggregation method is introduced. For each node, the features of neighboring nodes are aggregated using aggregation functions of different scales. These aggregated features are then merged or weighted to form the final feature representation of the node. The GraphSAGE model is trained using historical distribution network topology data feature data to learn node representations and capture normal distribution network topology data features. In the feature update step of each node, the features obtained by multi-scale aggregation are used to update the node representation.

[0009] Furthermore, aggregation functions of different scales are used to aggregate the features of neighbor nodes, including direct neighbors, second-degree neighbors, and third-degree neighbors; For each direct neighbor node , get its direct neighbor node set, and add these direct neighbor nodes to the first-degree neighbor set ; Similarly, based on the first-degree neighbor set, obtain the second-degree neighbor set ; Get the third-degree neighbor set based on the second-degree neighbor set , and then calculate respectively: ; ;

[0010] These aggregated features are then merged into a new representation of node v at layer l: .

[0011] Furthermore, the measurement object file and measurement data of the abnormal node are obtained, and the abnormal type of the abnormal node is obtained based on the abnormal type identification model, as follows: According to the real-time anomaly detection results, abnormal nodes are marked and their node numbers or connection relationship information are recorded; Obtain the measurement object files and real-time measurement data of the abnormal node, identify it based on the abnormal type recognition model, and determine the specific abnormal type of the abnormal node according to the output results of the abnormal type recognition model.

[0012] Furthermore, the anomaly type recognition model is constructed based on the support vector machine, as follows: In SVM, the optimal hyperplane is found by maximizing the margin to separate data points of different categories, features are extracted from the measurement object archives and measurement data of historical abnormal nodes, and feature vectors are constructed as training data sets; Train the SVM model using the training dataset; The decision function of the SVM model is as follows: ; Where x is the input feature vector, is the number of training samples, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples, is the kernel function, b is the bias term; For the input feature vector; The optimization goal is:

[0013] Where C is the regularization parameter, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples.

[0014] Furthermore, based on the abnormal nodes and their abnormal types, combined with the important nodes and community structure, the severity of the abnormality is evaluated according to the impact of the abnormal nodes and their position in the network, and the severity of the abnormality is obtained, as follows: Use PBetweenness Centrality algorithm to calculate the importance index of the node ; ; in, Represents the number of shortest paths from node s to node t; Indicates the number of shortest paths from node s to node t that pass through node v; Analyze the influence range of the community where the abnormal node is located, and calculate the connection strength CI within the community and the connection relationship CE between communities: ; ; Where R is the set of nodes in the community, Represents the connection weight between node u and node v; Consider the connection relationship between abnormal nodes and other nodes and obtain the node degree , which is the number of connections with other nodes; Based on abnormal node set ,The severity level of the anomaly is obtained by comprehensively considering the importance of the node, the ,influence range of the community where it is located, and the connection ,relationship of the abnormal node; ; Among them, Y is the severity level, is the node importance, The connection relationship of abnormal nodes, , and is the weight coefficient.

[0015] Further, visualization is performed as follows: based on the internal connections and inter-community connections of the community where the abnormal node is located, a connection diagram of the community where the abnormal node is located is drawn, and based on the calculated abnormal severity level, the level of each node is used as the attribute of the node; network visualization tools are used to display the connection relationship of the community where the abnormal node is located; based on the connection relationship between the abnormal node and other nodes connected to it, a connection relationship diagram of the abnormal node is drawn, and based on the position of the abnormal node in the network, a position diagram of the abnormal node is drawn.

[0016] The present invention has the following beneficial effects: 1. The present invention can effectively improve the accuracy and efficiency of distribution network topology anomaly identification and improve the operation reliability and safety of the distribution network through systematic topology construction, measurement data association, topology analysis, real-time anomaly detection and anomaly type identification; 2. The present invention abstracts power plants, substations, distribution lines, distribution transformers and users as nodes and uses unique identifiers to represent them, thereby ensuring the standardization and consistency of the topological structure, and defining branches according to the connection relationship between devices, clearly representing the power transmission path between devices, and facilitating the accurate construction of the topological structure of the distribution network; 3. The present invention adopts a depth-first search method to traverse the topological structure of the distribution network, and combines centrality analysis and community discovery methods to efficiently obtain the structure and characteristics of the distribution network, identify important nodes and community structures, and help understand the key nodes and local structures of the distribution network. By collecting real-time distribution network topological data features and performing detection based on anomaly detection models, anomalies in the topological structure can be discovered in a timely manner. Based on anomaly type recognition models, the anomaly types of abnormal nodes can be accurately identified. 4. The present invention is based on abnormal nodes and their abnormal types, combined with important nodes and community structures, and evaluates the severity of abnormalities according to the degree of influence of abnormal nodes and their positions in the network. It can more comprehensively understand the impact of abnormalities on the distribution network and ensure the pertinence and effectiveness of treatment measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0018] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments: refer to Figure 1 In this embodiment, a method for intelligently identifying abnormal topology of a distribution network is provided, comprising the following steps: S1: Abstract power plants, substations, distribution lines, distribution transformers, and users as nodes, which are represented by unique identifiers. According to the connection relationship between devices, branches are defined to represent the power transmission path between devices. According to the connection relationship between nodes and branches, the topological structure of the distribution network is constructed to form a connection relationship diagram between nodes. S2: Construct the association relationship between the measurement object files and measurement data of each node in the distribution network; S3: Based on the topological structure of the distribution network, the graph is traversed by the depth-first search method, and the structure and characteristics of the distribution network are obtained by combining centrality analysis and community discovery methods, and important nodes and community structures are identified; S4: Collect the features of real-time distribution network topology data, and perform detection based on the anomaly detection model, perform anomaly detection on the distribution network topology data, and obtain abnormal nodes; S5: Obtain the measurement object file and measurement data of the abnormal node, and obtain the abnormal type of the abnormal node based on the abnormal type recognition model; S6: Based on the abnormal nodes and their abnormal types, combined with important nodes and community structures, the severity of the abnormality is evaluated according to the impact of the abnormal nodes and their position in the network, the severity of the abnormality is obtained, and it is visualized.

[0019] In this embodiment, S2 is specifically: the measurement object file includes the device information, location, and parameter settings of the node, and is associated with the measurement data through a unique identifier; each measurement object file corresponds to real-time or historically collected measurement data, which is used to describe the device status and power system operation status; the physical connection relationship between the devices in the measurement object file is described through the topological connection relationship (such as the transmission line connecting the power plant and the substation); the measurement data is associated through the topological connection relationship in the measurement object file.

[0020] In this embodiment, S3 is specifically: Starting from a starting node, recursively visit its neighbor nodes and mark the visited nodes until all nodes have been visited; Calculate the degree centrality, betweenness centrality and closeness centrality of nodes and identify important nodes; ; ;

[0021] in, is the degree centrality of node v; is the degree of node v, that is, the number of edges directly connecting node v to other nodes; N is the total number of nodes; represents betweenness centrality; Represents the number of shortest paths from node s to node t; represents the number of shortest paths passing through node v; represents the closeness centrality, Represents the shortest path length from node v to node u; Using modularity algorithm for community discovery to identify community structures in distribution networks; ; Among them, Q is the modularity, represents the connection weight between node u and node v, , They represent the degrees of nodes u and v respectively, m represents the total number of edges, is an indicator function, which is 1 if node u and node v belong to the same community, otherwise it is 0.

[0022] In this embodiment, the anomaly detection model is built based on the GraphSAGE model, as follows: Build a GraphSAGE model to learn node representations and capture the relationships between nodes. Update the node representation by sampling neighbor node features and aggregating them. In the GraphSAGE model, the representation update of node v is calculated as follows: ; ; in, is the aggregation result of the representation of the neighbor nodes of node v at the lth layer; is the representation of node v at layer l; is the set of neighbor nodes of node v; is the neighbor node feature aggregation function, specifically ; is the weight matrix of the lth layer; is the activation function; CONCAT is the concatenation function, which means that the node v is The representation of the layer is concatenated with the representation of the neighbor nodes of node v at layer l; In the feature aggregation stage, a multi-scale feature aggregation method is introduced. For each node, the features of neighboring nodes are aggregated using aggregation functions of different scales. These aggregated features are then merged or weighted to form the final feature representation of the node. The GraphSAGE model is trained using historical distribution network topology data feature data to learn node representations and capture normal distribution network topology data features. In the feature update step of each node, the features obtained by multi-scale aggregation are used to update the node representation.

[0023] In this embodiment, aggregation functions of different scales are used to aggregate features of neighbor nodes, including direct neighbors, second-degree neighbors, and third-degree neighbors; For each direct neighbor node , get its direct neighbor node set, and add these direct neighbor nodes to the first-degree neighbor set ; Similarly, based on the first-degree neighbor set, obtain the second-degree neighbor set ; Get the third-degree neighbor set based on the second-degree neighbor set , and then calculate respectively: ; ;

[0024] These aggregated features are then merged into a new representation of node v at layer l: .

[0025] In this embodiment, the measurement object file and measurement data of the abnormal node are obtained, and the abnormal type of the abnormal node is obtained based on the abnormal type identification model, as follows: According to the real-time anomaly detection results, abnormal nodes are marked and their node numbers or connection relationship information are recorded; Obtain the measurement object files and real-time measurement data of the abnormal node, identify it based on the abnormal type recognition model, and determine the specific abnormal type of the abnormal node according to the output results of the abnormal type recognition model.

[0026] In this embodiment, the abnormal type recognition model is constructed based on a support vector machine, as follows: In SVM, the optimal hyperplane is found by maximizing the margin to separate data points of different categories, features are extracted from the measurement object archives and measurement data of historical abnormal nodes, and feature vectors are constructed as training data sets; Use the training data set to train the SVM model and adjust the hyperparameters (such as kernel function type, regularization parameters, etc.) to obtain the best classification effect; The decision function of the SVM model is as follows: ; Where x is the input feature vector, is the number of training samples, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples, is the kernel function (using linear kernel function or RBF kernel), b is the bias term; For the input feature vector; The optimization goal is:

[0027] Where C is the regularization parameter, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples.

[0028] In this embodiment, based on the abnormal nodes and their abnormal types, combined with important nodes and community structures, the severity level of the abnormality is evaluated according to the impact of the abnormal nodes and their positions in the network, and the severity level of the abnormality is obtained, as follows: Use PBetweenness Centrality algorithm to calculate the importance index of the node ; ; in, Represents the number of shortest paths from node s to node t; Indicates the number of shortest paths from node s to node t that pass through node v; Analyze the influence range of the community where the abnormal node is located, and calculate the connection strength CI within the community and the connection relationship CE between communities: ; ; Where R is the set of nodes in the community, Represents the connection weight between node u and node v; Consider the connection relationship between abnormal nodes and other nodes and obtain the node degree , which is the number of connections with other nodes; Based on abnormal node set ,The severity level of the anomaly is obtained by comprehensively considering the importance of the node, the ,influence range of the community where it is located, and the connection ,relationship of the abnormal node; ; Among them, Y is the severity level, is the node importance, The connection relationship of abnormal nodes, , and is the weight coefficient.

[0029] In this embodiment, visualization is specifically as follows: based on the internal connections and inter-community connections of the community where the abnormal node is located, a connection diagram of the community where the abnormal node is located is drawn, and based on the calculated abnormal severity level, the level of each node is used as the attribute of the node; a network visualization tool is used to display the connection relationship of the community where the abnormal node is located; based on the connection relationship between the abnormal node and other nodes connected to it, a connection relationship diagram of the abnormal node is drawn, and based on the position of the abnormal node in the network, a position diagram of the abnormal node is drawn It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0030] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0031] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0032] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0033] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.

Claims

1. A method for intelligently identifying abnormal topology of a distribution network, characterized in that: The following steps are involved: S1: Abstract power plants, substations, distribution lines, distribution transformers, and users as nodes, which are represented by unique identifiers. According to the connection relationship between devices, branches are defined to represent the power transmission path between devices. According to the connection relationship between nodes and branches, the topological structure of the distribution network is constructed to form a connection relationship diagram between nodes. S2: Construct the association relationship between the measurement object files and measurement data of each node in the distribution network; S3: Based on the topological structure of the distribution network, the graph is traversed by the depth-first search method, and the structure and characteristics of the distribution network are obtained by combining centrality analysis and community discovery methods, and important nodes and community structures are identified; S4: Collect the features of real-time distribution network topology data, and perform detection based on the anomaly detection model, perform anomaly detection on the distribution network topology data, and obtain abnormal nodes; S5: Obtain the measurement object file and measurement data of the abnormal node, and obtain the abnormal type of the abnormal node based on the abnormal type recognition model; S6: Based on the abnormal nodes and their abnormal types, combined with important nodes and community structures, the severity of the abnormality is evaluated according to the impact of the abnormal nodes and their position in the network, the severity of the abnormality is obtained, and it is visualized.

2. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: Specifically, S2 is as follows: the measurement object file includes the device information, location, and parameter settings of the node, and is associated with the measurement data through a unique identifier; each measurement object file corresponds to the measurement data collected in real time or historically, and is used to describe the device status and the operation of the power system; the physical connection relationship between the devices in the measurement object file is described through the topological connection relationship; the measurement data is associated through the topological connection relationship in the measurement object file.

3. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: The S3 is specifically: Starting from a starting node, recursively visit its neighbor nodes and mark the visited nodes until all nodes have been visited; Calculate the degree centrality, betweenness centrality and closeness centrality of nodes and identify important nodes; ; ； ； in, is the degree centrality of node v; is the degree of node v, that is, the number of edges directly connecting node v to other nodes; N is the total number of nodes; represents betweenness centrality; Represents the number of shortest paths from node s to node t; represents the number of shortest paths passing through node v; represents the closeness centrality, Represents the shortest path length from node v to node u; Using modularity algorithm for community discovery to identify community structures in distribution networks; ； Where Q is the modularity, represents the connection weight between node u and node v, , They represent the degrees of nodes u and v respectively, m represents the total number of edges, is an indicator function, which is 1 if node u and node v belong to the same community, otherwise it is 0.

4. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: The anomaly detection model is built based on the GraphSAGE model, as follows: Build a GraphSAGE model to learn node representations and capture the relationships between nodes. Update the node representation by sampling neighbor node features and aggregating them. In the GraphSAGE model, the representation update of node v is calculated as follows: ; ; in, is the aggregation result of the representation of the neighbor nodes of node v at the lth layer; is the representation of node v at layer l; is the set of neighbor nodes of node v; is the neighbor node feature aggregation function, specifically ; is the weight matrix of the lth layer; is the activation function; CONCAT is the concatenation function, which means that the node v is The representation of the layer is concatenated with the representation of the neighbor nodes of node v at layer l; In the feature aggregation stage, a multi-scale feature aggregation method is introduced. For each node, the features of neighboring nodes are aggregated using aggregation functions of different scales. Then, these aggregated features are merged or weighted to form the final feature representation of the node. The GraphSAGE model is trained using historical distribution network topology data feature data to learn node representations and capture normal distribution network topology data features. In the feature update step of each node, the features obtained by multi-scale aggregation are used to update the node representation.

5. A method for intelligently identifying distribution network topology anomalies according to claim 4, characterized in that: Said Aggregate the features of neighbor nodes using aggregation functions of different scales, including direct neighbors, second-degree neighbors, and third-degree neighbors; For each direct neighbor node , get its direct neighbor node set, and add these direct neighbor nodes to the first-degree neighbor set ; Similarly, based on the first-degree neighbor set, obtain the second-degree neighbor set ; Get the third-degree neighbor set based on the second-degree neighbor set , and then calculate respectively: ; ； ； These aggregated features are then merged into a new representation of node v at layer l: 。 6. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: The process of obtaining the measurement object file and measurement data of the abnormal node and obtaining the abnormal type of the abnormal node based on the abnormal type identification model is as follows: According to the real-time anomaly detection results, abnormal nodes are marked and their node numbers or connection relationship information are recorded; Obtain the measurement object files and real-time measurement data of the abnormal node, identify it based on the abnormal type recognition model, and determine the specific abnormal type of the abnormal node according to the output results of the abnormal type recognition model.

7. A method for intelligently identifying distribution network topology anomalies according to claim 6, characterized in that: The abnormal type recognition model is built based on support vector machine, as follows: In SVM, the optimal hyperplane is found by maximizing the margin to separate data points of different categories, features are extracted from the measurement object archives and measurement data of historical abnormal nodes, and feature vectors are constructed as training data sets; Train the SVM model using the training dataset; The decision function of the SVM model is as follows: ； Where x is the input feature vector, is the number of training samples, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples, is the kernel function, b is the bias term; For the input feature vector; The optimization goal is: Where C is the regularization parameter, It is The coefficients of the support vectors corresponding to the input eigenvectors, It is The input feature vectors correspond to the labels of the samples.

8. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: Based on the abnormal nodes and their abnormal types, combined with important nodes and community structures, the severity of the abnormality is evaluated according to the impact of the abnormal nodes and their positions in the network, and the severity of the abnormality is obtained, as follows: Use PBetweenness Centrality algorithm to calculate the importance index of the node ; ； in, Represents the number of shortest paths from node s to node t; Indicates the number of shortest paths from node s to node t that pass through node v; Analyze the influence range of the community where the abnormal node is located, and calculate the connection strength CI within the community and the connection relationship CE between communities: ； ； Where R is the set of nodes in the community, Represents the connection weight between node u and node v; Consider the connection relationship between abnormal nodes and other nodes and obtain the node degree , which is the number of connections with other nodes; Based on abnormal node set ,The severity level of the anomaly is obtained by comprehensively considering the importance of the node, the ,influence range of the community where it is located, and the connection ,relationship of the abnormal node; ； Among them, Y is the severity level, is the node importance, The connection relationship of abnormal nodes, , and is the weight coefficient.

9. The method for intelligently identifying distribution network topology anomalies according to claim 1, characterized in that: The visualization is specifically as follows: based on the internal connections and inter-community connections of the community where the abnormal node is located, a connection diagram of the community where the abnormal node is located is drawn, and based on the calculated abnormal severity level, the level of each node is used as the attribute of the node; using a network visualization tool to display the connection relationship of the community where the abnormal node is located; based on the connection relationship between the abnormal node and other nodes connected to it, a connection relationship diagram of the abnormal node is drawn, and based on the position of the abnormal node in the network, a position diagram of the abnormal node is drawn.

Images