Fault processing method and device based on main distribution network cooperation

Abstract

This invention discloses a fault handling method and apparatus based on main and distribution network coordination, belonging to the field of power system automation technology. The method includes: using a graph neural network to perform topology identification and fault location of the power grid, generating fault impact features; based on the spatial range and electrical connection relationships of the fault impact features, obtaining the switches at both ends of the faulty section of the power grid, and performing fault isolation through remote control tripping; analyzing the fault isolation information using topology impact analysis; and, combined with changes in the current topology of the power grid, generating a power grid protection strategy migration mechanism and initiating a fault recovery process to restore power supply. This invention achieves dynamic modeling of the power grid topology and accurate identification of faulty sections by using a graph neural network to embed the power grid graph and combining multi-channel time series information to generate node potential state vectors, thus improving the real-time performance and accuracy of fault location.

Description

Technical Field

[0001] This invention relates to the field of power system automation technology, and in particular to fault handling methods and devices based on the coordination of main and distribution networks. Background Technology

[0002] In the field of power system operation and control, the main distribution network, as a critical link in the power transmission and distribution process, plays a vital role in ensuring the reliability of power supply. Traditional fault handling methods are typically based on centralized data acquisition and analysis mechanisms. By monitoring information such as voltage parameters, current parameters, and switch status, and combining this with topology identification technology, fault segment identification and isolation operations are achieved. These methods rely on the preprocessing and integration of multi-source data and utilize power flow analysis and path tracing to locate faults and assess their impact range. Building upon this foundation, further employing remote tripping and load transfer strategies to implement fault isolation and power restoration constitutes the main technical approach for current power grid fault handling.

[0003] However, existing methods have certain limitations in dealing with complex operating conditions and the coordination of multi-source heterogeneous data. On the one hand, conventional topology identification methods mainly rely on static model construction, which makes it difficult to dynamically integrate real-time operating data, resulting in inaccurate extraction of fault impact features. On the other hand, in the process of generating power grid protection strategy migration mechanisms, there is a lack of dynamic adjustment capability for protection coordination relationships under the new topology, which limits the improvement of fault recovery efficiency and self-healing capability. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a fault handling method based on main and distribution network coordination to solve the problems of insufficient dynamics in existing power grid topology identification and weak adaptability of protection strategy migration mechanism after fault isolation.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a fault handling method based on main and distribution network coordination, comprising: collecting and preprocessing power grid operation status data to generate a power grid operation data stream; using a graph neural network to perform topology identification and fault location on the power grid based on the power grid operation data stream, generating fault impact features; obtaining the switches at both ends of the fault section of the power grid according to the spatial range and electrical connection relationship of the fault impact features, and performing fault isolation to generate fault isolation information; analyzing the fault isolation information using topology impact analysis, and combining the changes in the current topology of the power grid to generate a power grid protection strategy migration mechanism and initiate a fault recovery process to restore power supply.

[0008] As a preferred embodiment of the fault handling method based on primary and secondary network coordination described in this invention, wherein:

[0009] The power grid operation status data includes voltage parameters, current parameters, power parameters, switch status information, and protection device operation information;

[0010] The preprocessing includes filtering and denoising, handling missing values, correcting outliers, and data standardization.

[0011] As a preferred embodiment of the fault handling method based on main and distribution network coordination described in this invention, the generation of power grid operation data stream integrates the preprocessed power grid operation status data by using a multi-source heterogeneous data fusion method.

[0012] As a preferred embodiment of the fault handling method based on primary and distribution network coordination described in this invention, the specific steps for generating the power grid diagram are as follows:

[0013] Extract the electrical parameters of each node in the power grid operation data stream, identify the equipment types in the power grid, and generate graph structure nodes;

[0014] The topology connection data of the power grid is parsed using graph modeling to generate graph structure edges;

[0015] By integrating the nodes and edges of the graph structure, a power grid graph is generated.

[0016] As a preferred embodiment of the fault handling method based on main and distribution network coordination described in this invention, the specific steps of using a graph neural network to perform topology identification and fault location of the power grid are as follows:

[0017] The graph neural network is used to embed the power grid graph and combine it with multi-channel time series information in the power grid operation data stream to generate node potential state vectors.

[0018] Based on the node potential state vector, the regional division and line-transformer relationship in the power grid are identified, and the power grid topology is output by combining the electrical parameters in the power grid operation data stream.

[0019] The connection relationships in the power grid topology are analyzed to generate power grid fault sections.

[0020] As a preferred embodiment of the fault handling method based on main and distribution network coordination described in this invention, the generation of fault impact features is achieved by integrating the power grid topology and the faulty section of the power grid.

[0021] As a preferred embodiment of the fault handling method based on main and distribution network coordination described in this invention, the specific steps for obtaining the switches at both ends of the faulty section of the power grid based on the spatial range and electrical connection relationship of the fault impact characteristics are as follows:

[0022] Extract the edge set from the power grid diagram and generate electrical connection relationships;

[0023] The topology path tracing method is used to analyze the propagation path of the spatial range of fault impact characteristics, and adjacency matching is performed in combination with electrical connection relationships to generate the fault impact area.

[0024] Identify the boundary points between the fault-affected area and the normal operating area, and obtain the boundary point switches through the switch status evaluation method to generate a switch set.

[0025] As a preferred embodiment of the fault handling method based on primary and secondary network coordination described in this invention, the specific steps of performing fault isolation and generating fault isolation information are as follows:

[0026] The switch set is tripped, a remote tripping command is generated, and a safety verification is performed.

[0027] The automated master station executes electrical isolation operations on remote tripping commands to generate fault isolation information.

[0028] As a preferred embodiment of the fault handling method based on main and distribution network coordination described in this invention, the steps include: analyzing fault isolation information using topology impact analysis, combining the changes in the current power grid topology, generating a power grid protection strategy migration mechanism, and initiating a fault recovery process to restore power supply.

[0029] Based on fault isolation information, the changes in the current topology of the power grid are obtained, and the line load capacity in the fault-affected area is verified by combining power flow calculation method, and an isolation response analysis set is generated.

[0030] Perform protection strategy generation and logic update operations on the isolation response analysis set to generate a power grid protection strategy migration mechanism;

[0031] Based on the isolation response analysis set and the power grid protection strategy migration mechanism, a power supply restoration path is selected, and the fault recovery process is initiated.

[0032] Secondly, the present invention provides a fault handling device based on main and distribution network coordination, including a data acquisition module for acquiring power grid operation status data and performing preprocessing to generate power grid operation data stream;

[0033] The identification and location module is used to perform topology identification and fault location of the power grid based on the power grid operation data stream and using a graph neural network to generate fault impact features.

[0034] The fault isolation module is used to obtain the switches at both ends of the faulty section of the power grid based on the spatial range and electrical connection relationship of the fault impact characteristics, and to perform fault isolation and generate fault isolation information.

[0035] The startup analysis module is used to analyze fault isolation information using topology impact analysis. Combined with changes in the current power grid topology, it generates a power grid protection strategy migration mechanism and initiates a fault recovery process to restore power supply.

[0036] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the fault handling method based on main distribution network coordination as described in the first aspect of the present invention.

[0037] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the fault handling method based on primary and secondary network coordination as described in the first aspect of the present invention.

[0038] The beneficial effects are as follows: By using graph neural networks to embed and learn the power grid graph, and combining multi-channel time series information to generate node potential state vectors, dynamic modeling of the power grid topology and accurate identification of fault sections are achieved, improving the real-time performance and accuracy of fault location, and providing a high-quality data foundation for subsequent fault impact range analysis; by dynamically adjusting and logically updating the isolation response analysis set, based on the line load capacity and power supply path changes under the new topology, the protection coordination relationship is reconstructed and a power grid protection strategy migration mechanism is generated, realizing adaptive switching of protection logic after fault isolation, and improving stability and recovery efficiency under abnormal operating conditions. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart of a fault handling method based on primary and secondary network coordination.

[0041] Figure 2 The flowchart shows the main and distribution network collaborative fault handling architecture.

[0042] Figure 3 This is a flowchart for fault handling.

[0043] Figure 4 A flowchart detailing the power grid diagram modeling and fault location process. Detailed Implementation

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0047] Reference Figures 1-4 This is one embodiment of the present invention, which provides a fault handling method based on primary and secondary network coordination, including the following steps:

[0048] S1. Collect and preprocess power grid operation status data to generate a power grid operation data stream;

[0049] Power grid operation status data includes voltage parameters, current parameters, power parameters, switch status information, and protection device operation information;

[0050] It should be noted that voltage parameters are measured by voltage transformers to reflect the voltage levels of each node at different times, such as bus voltage or feeder voltage; current parameters are collected by current transformers to characterize the current flow in lines or equipment, such as the real-time current value of a 10 kV line; power parameters are obtained through synchronous measurement of voltage and current, including active power and reactive power, and are used to assess the energy transmission status of different areas in the power grid; switch status information is obtained through remote signaling signals to record the open and closed states of equipment such as circuit breakers and disconnectors, such as the main switch of a certain line being in the closed state; protection device action information comes from the action records of relay protection equipment, including protection start-up, trip output, and triggering of related fault criteria, such as the overcurrent protection action and the drive of the switch to trip after a phase-to-phase short circuit occurs on a certain line.

[0051] Preprocessing includes filtering and denoising, handling missing values, correcting outliers, and data standardization.

[0052] Specifically, the sliding window averaging method is used to suppress high-frequency noise in voltage, current, and power parameters. For example, it smooths fluctuations caused by electromagnetic interference in the current measurement sequence of a certain line. Missing value processing supplements missing voltage or power parameters by filling in values ​​from adjacent time points. For example, when active power data is missing at a certain time point, the average value of the previous and next time points is used as a substitute. Outlier correction identifies and corrects voltage, current, and protection device operation information that exceed the rated range of the equipment. For example, when the current measurement value of a certain line exceeds the upper limit of the transformer range, it is judged as an anomaly and replaced with the historical average value. The rated range of equipment refers to the maximum and minimum values ​​of voltage, current, or power that various electrical equipment in the power system can withstand under normal operating conditions. Data standardization uses the min-max normalization method to normalize voltage, current, and power parameters.

[0053] The preprocessed power grid operation status data is integrated using a multi-source heterogeneous data fusion method to generate a power grid operation data stream.

[0054] Furthermore, a time synchronization method is used to perform time-stamp correction on voltage parameters, current parameters, and protection device action information from different acquisition nodes based on a unified time reference. For example, the voltage parameters, current parameters, and protection device action information collected by the substation bay level equipment and the dispatching main control unit are unified to the same millisecond-level time series. The data alignment method aligns each parameter sequence at a fixed time interval and interpolates power parameters and switch status information for different sampling frequencies to ensure that power parameters and switch status information remain consistent in the time dimension. The feature fusion method combines voltage parameters, current parameters, and power parameters under the same timestamp into a complete operating status description, and combines switch status information and protection device action information to construct a complete set of power grid operating status descriptions, generating a power grid operating data stream.

[0055] S2. Based on the power grid operation data stream, use graph neural networks to perform topology identification and fault location of the power grid, and generate fault impact features;

[0056] The electrical parameters of each node in the power grid operation data stream are extracted using the sliding window feature extraction method, and the equipment type in the power grid is identified by the equipment label matching method to generate graph structure nodes.

[0057] Specifically, the voltage, current, and power parameter sequences are segmented using a fixed time length. For example, a time window with a length of 10 seconds and a step size of 2 seconds is used, and the mean, variance, and peak value within each window are calculated as the nodal electrical parameter features. The fixed time length refers to the uniform time period set when segmenting time series data. For example, when using the sliding window feature extraction method in the power grid operation data stream, continuous voltage, current, and power parameters are segmented according to the same time interval in order to extract local features.

[0058] The collected switch status information is compared with a set of equipment type labels using a device label matching method. For example, the unique identifiers of equipment such as circuit breakers, disconnectors, and transformers are used to identify equipment types, thereby determining the type of equipment connected to each node in the power grid. The equipment type label set refers to the combination of labels used to identify different types of power equipment, which is set before power grid data processing. This set is compared with the collected switch status information in the device label matching method to identify the specific type of equipment connected to each node in the power grid. Combining the extracted electrical parameter features with the identified equipment type information, a graph structure node containing node attributes and connection relationships is constructed, generating the graph structure node.

[0059] The topology connection data of the power grid is parsed using graph modeling to generate graph structure edges;

[0060] It should be noted that the physical connection relationship between each node is determined based on the topology connection data of the power grid. The topology connection data of the power grid refers to the data set describing the physical connection relationship between each node in the power grid, including the connection status between nodes, the type of connected equipment, and the corresponding electrical connection attributes. The physical connection relationship is determined based on the substation wiring diagram, equipment connection record, and switch status information. The node number connected to different equipment is identified, and the association between equipment and nodes is established.

[0061] Create edges between corresponding nodes to represent physical connections; assign weights to each edge based on the electrical properties of the connected devices, using the electrical properties corresponding to the resistance values ​​of the devices as the weights of the corresponding edges in the graph structure; for example, if the resistance value of a certain line segment is 0.5 ohms, then this value is used as the weight of the edge; generate the graph structure edges through the above process.

[0062] The adjacency matrix method is used to integrate the nodes and edges of the graph structure to generate a power grid graph.

[0063] Furthermore, the total number of all graph-structured nodes in the power grid is counted and used as the number of rows and columns of the adjacency matrix to construct a weighted undirected graph adjacency matrix. For example, if there are 5 graph-structured nodes in the power grid, a 5x5 matrix is ​​constructed. The connection relationship between nodes represented by each graph-structured edge is mapped using a matrix. For example, when there is an edge between nodes with identifiers 2 and 3, the corresponding edge weight value is filled in the 2nd row and 3rd column and the 3rd row and 2nd column of the matrix. If there is no graph-structured edge between two nodes, 0 is filled in the corresponding position to indicate no direct connection. Finally, the connection relationship and weight between all graph-structured nodes are completely expressed through the matrix, forming the power grid graph.

[0064] The graph neural network is used to embed the power grid graph and combine it with multi-channel time series information in the power grid operation data stream to generate node potential state vectors.

[0065] Specifically, the power grid graph is represented in the form of an adjacency matrix and a node feature matrix. The adjacency matrix records the connection relationships between nodes described by the graph structure edges and the corresponding edge weights. The node feature matrix consists of electrical parameters extracted from the graph structure nodes, including multi-channel time series information such as voltage parameters, current parameters, and power parameters.

[0066] In each layer of a graph neural network, each node aggregates feature information from its neighbors, combines it with its own current features, performs a linear transformation using a trainable parameter matrix, and updates its node representation using a non-linear activation function. Combining its own current features means that during the node representation update process, the existing feature information of the current node is fused with the information aggregated from its neighbors to generate a more representative node representation. The trainable parameter matrix is ​​a set of parameters randomly initialized during the construction of the graph neural network and continuously updated and optimized during training, used to perform linear transformations on node features. For example, in the first layer, the node with identifier 2 collects feature information from nodes with identifiers 1 and 3, and calculates a new node representation by combining it with its own features. The process propagates layer by layer through multiple graph neural network layers, enabling each node to ultimately obtain a comprehensive representation that includes its local topological role in the power grid graph and the dynamic changes of multi-channel time series. In the first layer, each node only aggregates the features of its direct neighbors to generate a preliminary node representation. In the second layer, the node further aggregates the information of its neighbors based on the output of the first layer, thereby capturing the topological and operational state information within a two-hop range of itself. In subsequent network layers such as the third and fourth layers, the node gradually integrates structural and time series information from a wider neighborhood, ultimately generating the node's latent state vector.

[0067] Based on the node potential state vector, spectral clustering is used to identify the regional division and line-transformation relationship in the power grid. Combined with the electrical parameters in the power grid operation data stream, the power grid topology is output.

[0068] It should be noted that the potential state vectors of all nodes in the power grid diagram are used to form a feature matrix. For example, each potential state vector is a numerical vector of length 128. A similarity matrix is ​​constructed, and the degree of electrical coupling is reflected by calculating the cosine similarity of the feature vectors between each pair of nodes. The similarity matrix is ​​normalized and the corresponding Laplace matrix is ​​solved.

[0069] Calculate the cosine similarity of the feature vectors between each pair of nodes. The expression reflecting the degree of electrical coupling is:

[0070] ;

[0071] in, This represents the mathematical index of any selected graph node in the power grid diagram, used to identify the first object participating in the electrical coupling analysis; Indicates another independent entity in the power grid diagram. The mathematical index of the graph structure node is used to identify the second object involved in the comparison; This indicates the first object participating in the electrical coupling analysis; Indicates and The second object to be compared; Representing vectors and The dot product;

[0072] The Laplace matrix is ​​decomposed to form a new low-dimensional feature matrix. Based on this, the K-means clustering algorithm is applied to divide all nodes into k clusters. For example, setting k=4 can divide 100 nodes in the power grid into four regions with similar topological roles and electrical characteristics, enabling the identification of different regions in the power grid. Simultaneously, the connection patterns between regions are analyzed based on the clustering results of the nodes, identifying the affiliation between lines and transformers, forming a line-transformer relationship description. The line-transformer relationship refers to the connection affiliation between lines and transformers in the power grid, describing which lines are connected to the same transformer, which transformers are responsible for supplying power to specific regions, and other structural characteristics, reflecting the hierarchical organization of the power grid in physical connections and energy transmission. Combining voltage, current, and power parameters from the power grid operation data stream, the line-transformer relationship is used to further verify the consistency of electrical parameters within the regions, confirming the rationality of the region division and line-transformer relationship, and finally outputting the power grid topology.

[0073] Power flow analysis is used to analyze the connection relationships in the power grid topology and generate power grid fault sections.

[0074] Furthermore, based on the nodes, graph edges, and corresponding edge weights described in the power grid topology, a power grid model including node admittance matrices and power injection assumptions is established. According to the graph nodes and edges described in the power grid topology, the connection relationships between all nodes in the power grid and the electrical parameters corresponding to each edge are determined. These electrical parameters are then converted into elements in the node admittance matrix. The node admittance matrix is ​​a square matrix with the number of rows and columns equal to the number of graph nodes. Each off-diagonal element represents the mutual admittance between two nodes, calculated from the admittance value of the corresponding graph edge. The diagonal elements represent the self-admittance of each node, composed of the mutual admittance between the node and all other connected nodes and the grounding admittance of the equipment connected to the node. Simultaneously, an initial active and reactive power injection value is set for each graph node. For example, the node with identifier 1 is designated as the balancing node, with its active power set to a known value of 100 MW and reactive power set to 50 Mvar. The power injection values ​​for other nodes can be set based on load forecasting or historical data, thus constructing a complete power grid model.

[0075] Based on the node admittance matrix and power injection assumptions in the power grid model, a reference node and a slack node are set. The Newton-Raphson method is used to solve the nonlinear power flow equations, calculate the voltage magnitude and phase angle of each node, and the active and reactive power transmitted on each graph edge. For example, if the power transmitted by the line with identifier 3 is found to be much higher than the rated capacity during the calculation, it is determined that there is an abnormal operating state in the area where the line is located. Combined with the abnormal electrical characteristics in the power grid operation data stream, the sections in the power grid that may be faulty are located based on the above analysis, and the power grid fault sections are generated.

[0076] A graph traversal analysis method is used to integrate the power grid topology and fault sections to generate fault impact characteristics.

[0077] Specifically, a graph structure for traversal is constructed. Based on the node numbers and connections provided by the power grid topology, a traversal graph structure consisting of nodes and edges is created. For example, if there are 10 nodes in the power grid and they are interconnected by several lines, then the node numbers, such as 1 to 10, are used as nodes in the graph, and edges are established between the corresponding nodes based on the node numbers at both ends of each line. Each edge contains related electrical parameters, such as resistance and reactance values, which serve as edge weights. Taking the abnormal lines or transformers identified in the faulty sections of the power grid as starting points, a depth-first search algorithm is applied to visit adjacent nodes and edges sequentially along the edges of the traversal graph structure from the starting point, tracing the possible propagation paths of the fault. For example, if the line with identifier 3 is identified as a faulty section, the graph traversal starts from the nodes connected to the line to identify whether the directly connected nodes, lines, and transformers are affected by the fault.

[0078] By combining the number of nodes, the number of edges, and the description of line-to-line relationships involved in the traversal process, the range of the affected area is calculated, and the changing trends of voltage and current parameters of each node are recorded; finally, fault impact characteristics are generated.

[0079] S3. Based on the spatial range and electrical connection relationship of the fault impact characteristics, the switches at both ends of the fault section of the power grid are obtained by using the topology path analysis method, and the fault isolation is performed by the remote control tripping method to generate fault isolation information.

[0080] The SCADA analytical method is used to extract the edge set in the power grid diagram and generate electrical connection relationships;

[0081] It should be noted that the connection information between all nodes in the power grid diagram is obtained. This connection information comes from telemetry, tele-signaling signals, and switch status information and protection device action information identified by equipment tag matching. Based on the definition of graph structure edges described in the power grid diagram, each edge representing the physical connection relationship between nodes is extracted and the edges are grouped into an edge set. For example, in a certain analysis, if there is an edge connected by a circuit breaker between nodes with identifiers 2 and 3, then the edge is included in the edge set. The definition of a graph structure edge includes that the two ends of the edge must connect two graph structure nodes and represent actual existing electrical equipment or physical lines, such as transmission lines, transformers, circuit breakers, etc. Each edge also has corresponding electrical attributes, such as resistance value, reactance value, or rated capacity, and is accompanied by equipment status information, such as the open / closed state of a switch.

[0082] Record the electrical connection attributes corresponding to each edge, such as line type, equipment number, and rated parameters; finally, generate the electrical connection relationship.

[0083] The topology path tracing method is used to analyze the propagation path of the spatial range of fault impact characteristics, and adjacency matching is performed in combination with electrical connection relationships to generate the fault impact area.

[0084] Furthermore, based on the graph structure nodes and edges in the power grid diagram, the initial fault node or edge identified by the fault impact characteristics is used as the propagation starting point. A depth-first search strategy is adopted to sequentially visit adjacent nodes and edges along the graph structure edges, recording all possible paths for fault propagation. For example, if the line with identifier 3 is determined to be a faulty section, starting from the node connected to the line, adjacent lines and transformer equipment are traced layer by layer to identify all nodes and edges that may be affected by the fault. At the same time, the switch status information in the electrical connection relationship is combined to determine whether the path is conductive. If the circuit breaker on a certain path is in the open state, the path tracing is terminated. Finally, the fault impact area is generated based on all visited nodes and edges.

[0085] The boundary points between the fault-affected area and the normal operation area are identified by the adjacency matching method, and the boundary point switches are obtained by the switch state evaluation method to generate a switch set;

[0086] Specifically, all nodes in the power grid diagram are divided into nodes belonging to the fault-affected area and nodes belonging to the normal operation area. The initial fault node is identified based on the fault impact characteristics. Starting from the fault node, the topology path tracing method is used to visit all potentially affected adjacent nodes along the graph structure edges in the power grid diagram. For example, if the line with identifier 3 is determined to be a faulty section, then starting from the nodes connected at both ends of the line, the adjacent nodes are traced layer by layer until no new affected nodes are found. All nodes that are visited and have a fault propagation path are marked as nodes within the fault-affected area. Conversely, nodes that are not covered by the fault propagation path and whose electrical parameters, such as voltage and current, are within the normal range are classified as nodes within the normal operation area.

[0087] For each node within the fault-affected area, an adjacency analysis is performed to check whether all connected adjacent nodes include nodes belonging to the normal operating area. If a node belongs to the fault-affected area and has at least one adjacent node belonging to the normal operating area, the node is marked as a candidate boundary point node, and the graph structure edge between it and its adjacent nodes is recorded. For example, if a node with identifier 2 is determined to be a node in the fault-affected area and a node with identifier 5 is determined to be a node in the normal operating area, and if there is a graph structure edge between these two nodes and a circuit breaker device is configured there, then both nodes with identifiers 2 and 5 are confirmed as candidate boundary point nodes, and the corresponding circuit breaker is considered a potential boundary point switch. A circuit breaker device is an electrical device used to realize the automatic connection and disconnection control of circuits in the power grid. It can cut off or close current loops under normal or fault conditions, playing a protective and isolation role.

[0088] Based on the analysis of remote signaling signals, the actual open / closed state of the switching equipment connected to each edge of the diagram structure is determined. If the switch is closed, it indicates that the current path is connected and can be used as an effective isolation operation target; if the switch is open, it indicates that the path is disconnected and it will not participate in the current fault isolation operation. Remote signaling signals are digital signals used to reflect the status of switching equipment (such as circuit breakers, disconnectors, etc.), usually represented by "0" or "1" to indicate the open or closed state of the equipment. In distribution automation systems, remote signaling signals are collected by field terminal equipment (such as FTUs, DTUs) and uploaded to the dispatch automation master station system through communication channels for real-time monitoring of the operating status of power grid equipment. All boundary point switches that meet the conditions are grouped into a switch set.

[0089] The remote control tripping method is used to trip the switch set, generate remote tripping commands and perform safety verification;

[0090] It should be noted that, based on the graph structure nodes and edges in the power grid diagram, and combined with the boundary point identification results between the fault-affected area and the normal operation area, the remote control authority of each circuit breaker device in the switch set is confirmed. A remote tripping command is sent from the dispatch master station to each circuit breaker device belonging to the switch set, containing the unique identifier of the target circuit breaker device and the operation command; for example, a tripping command is sent to the circuit breaker device with the identifier CB-102. The dispatch master station is used to realize power grid operation monitoring, remote control, and fault handling. It is usually deployed in the local dispatch, distribution dispatch, or main control center, and has functions such as data acquisition, status monitoring, power flow calculation, protection control, and remote operation. The dispatch master station interacts with field equipment through the communication network, can obtain power grid operating parameters in real time, and can send tripping or closing commands to circuit breaker devices with remote control authority, realizing centralized management and rapid response of power grid equipment.

[0091] The system checks the current remote signaling status of the circuit breaker to determine if it is in a closed state. If it is in an open state, the operation procedure is skipped. If it is in a closed state, the system further checks if the current parameters of the nodes connected to the circuit breaker are below the current threshold to determine if the circuit breaker meets the safety conditions for remote tripping. The current threshold is set comprehensively based on the rated current parameters of the circuit breaker, power grid operation specifications, and safety requirements for fault handling. It is used to determine whether the safety conditions for remote tripping are met. The specific value range is usually 1.05 to 1.2 times the maximum load current that the circuit breaker is allowed to operate stably for a long time. This ensures that the tripping operation is performed under low load conditions, avoiding risks such as arcing, overload, or equipment damage caused by high current. For example, if the rated current of a circuit breaker is 500A, and the current threshold is set to 20% of the rated current, the corresponding value is 100A. Setting the current threshold to 10% of the rated current ensures that the tripping operation will not cause arcing or overload risks. After all safety verification conditions are met, the remote tripping operation is performed.

[0092] The automated master station executes electrical isolation operations on remote tripping commands to generate fault isolation information.

[0093] Furthermore, the dispatch automation master station, according to the communication protocol, sends remote tripping commands to each eligible circuit breaker in the switch set. The commands include the target device's unique identifier and the operation command. The communication protocol refers to the pre-agreed data transmission rules between the dispatch automation master station and the circuit breaker devices. It standardizes the format, transmission method, and response mechanism of the remote tripping commands, defining the arrangement of fields such as the command's start identifier, target device's unique identifier, operation command type, checksum, and terminator. It also specifies the data transmission rate, communication port, and error retransmission mechanism to ensure that the control commands sent by the dispatch automation master station can be correctly parsed and executed by the circuit breaker devices, while guaranteeing the stability of the communication process and the reliability of the operation.

[0094] Upon receiving a command, the circuit breaker executes a tripping action. The command content is parsed, and the unique identifier of the circuit breaker is verified to match, as well as the validity of the operation command. This is achieved by comparing the unique identifier of the target device contained in the remote tripping command with the identifier stored locally on the circuit breaker, and by checking whether the operation command is within the allowed instruction set. If the execution conditions are met, the internal control circuit is activated, driving the operating mechanism to separate the contacts, completing the tripping action. The execution result is then fed back to the dispatch automation master station via remote signaling. Finally, fault isolation information is generated based on the response status of all circuit breakers.

[0095] S4. Use topology impact analysis to analyze fault isolation information, combine it with the changes in the current power grid topology, generate a power grid protection strategy migration mechanism, and initiate a fault recovery process to restore power supply.

[0096] Based on fault isolation information, the topology path tracing method is used to obtain the changes in the current topology of the power grid, and the power flow calculation method is combined to verify the line load capacity in the fault-affected area and generate an isolation response analysis set.

[0097] Specifically, based on the graph structure nodes and edges described in the power grid diagram, and combined with the circuit breaker devices that have been tripped recorded in the fault isolation information, the corresponding graph structure edge status in the power grid diagram is updated. For example, after the circuit breaker device with the identifier CB-102 performs a tripping operation, the graph structure edge between the nodes with the identifiers 3 and 4 in the corresponding power grid diagram is marked as disconnected. A depth-first search strategy is used to perform path tracing on the updated power grid diagram to identify new power supply islands or undervoltage areas formed due to disconnection, and to obtain the changes in the current power grid topology.

[0098] By combining power flow calculation methods, power flow distribution calculations are performed on lines within the fault-affected area to verify whether the load capacity of each line exceeds its rated capacity. For example, if the rated capacity of a line is 100 MVA, and the power flow calculation results show that the actual transmission power is 95 MVA, then the line is determined to have the ability to continue operating. Finally, an isolation response analysis set is generated based on the changes in the power grid topology and the power flow verification results.

[0099] The power flow distribution within the fault-affected area is calculated using power flow calculation methods, and the expression is as follows:

[0100] ;

[0101] in, Represents a node in a power grid diagram. Representatives and nodes Another connected node, This indicates the total power transmitted over the line. Indicates from node The active power flowing to node k, Indicates from node The reactive power flowing to node k.

[0102] A topology-driven protection reconfiguration method is used to dynamically adjust and logically update the isolation response analysis set, generating a power grid protection strategy migration mechanism.

[0103] Furthermore, by analyzing the connectivity changes of graph nodes and edges described in the power grid diagram before and after fault isolation operations, the set of isolated nodes that cannot be reached from the power supply nodes due to circuit breaker tripping is identified, thereby determining the power supply path interruption area caused by the fault isolation operation. Combined with the line load capacity assessment results recorded in the isolation response analysis, the protection action logic of the affected nodes is reconfigured. For example, if the load rate of a line with the identifier L-205 increases from 65% to 92% after fault isolation, approaching the rated capacity limit, the backup protection action time limit of the line is dynamically shortened according to the protection setting rules. The protection setting rules refer to the technical specifications that are pre-set according to the power grid operation requirements and equipment parameters to determine the protection action type and action time limit that the circuit breaker equipment should perform under different operating conditions.

[0104] Based on the graph nodes and edges described in the power grid diagram, boundary nodes that remain operational after fault isolation operations and are directly connected to the fault-affected area are identified. The power supply source and power flow distribution of these boundary nodes in the new topology are analyzed using the topology path tracing method. The new topology originates from the updated power grid diagram based on fault isolation information, reflecting changes in node and edge connections caused by circuit breaker tripping operations. Simultaneously, the protection action delay of critical lines is shortened to improve fault clearing speed and meet the speed requirement. Finally, the protection coordination relationship between adjacent nodes is updated to ensure the new protection strategy meets the selectivity and speed requirements. A power grid protection strategy migration mechanism is generated through the above adjustment and update process.

[0105] Based on the isolation response analysis set and the power grid protection strategy migration mechanism, the power supply restoration path is selected using the load transfer method, and the fault recovery process is initiated.

[0106] It should be noted that, based on the power grid diagram and the disconnected graph structure edges marked in the fault isolation information, the set of undervoltage nodes is identified; based on the updated protection coordination relationship in the power grid protection strategy migration mechanism, the available backup power supply paths for each undervoltage node are evaluated; for example, if there are two potential power supply paths for node N-15, which are connected to the normal operation area through lines identified by L-207 and L-209 respectively, then the path that meets the safety constraints and has the strongest power supply capacity is selected as the power supply restoration path according to the line load rate, rated capacity and protection action time limit.

[0107] After determining the recovery path, a corresponding circuit breaker closing operation sequence is generated, and a remote closing command is sent to the dispatch automation master station to start the fault recovery process and realize the power supply restoration to the lost voltage area.

[0108] This embodiment also provides a fault handling device based on primary and secondary network coordination, including:

[0109] The data acquisition module is used to collect power grid operation status data and preprocess it to generate a power grid operation data stream;

[0110] The identification and location module is used to perform topology identification and fault location of the power grid based on the power grid operation data stream and using a graph neural network to generate fault impact features.

[0111] The fault isolation module is used to obtain the switches at both ends of the faulty section of the power grid based on the spatial range and electrical connection relationship of the fault impact characteristics, and to perform fault isolation and generate fault isolation information.

[0112] The startup analysis module is used to analyze fault isolation information using topology impact analysis. Combined with changes in the current power grid topology, it generates a power grid protection strategy migration mechanism and initiates a fault recovery process to restore power supply.

[0113] This embodiment also provides a computer device applicable to the fault handling method based on main distribution network coordination, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the fault handling method based on main distribution network coordination as proposed in the above embodiment.

[0114] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0115] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the fault handling method based on main-distribution network coordination as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0116] In summary, this invention achieves dynamic modeling of the power grid topology and accurate identification of fault sections by using graph neural networks to embed and learn the power grid graph and combining multi-channel time series information to generate node potential state vectors. This improves the real-time performance and accuracy of fault location and provides a high-quality data foundation for subsequent fault impact range analysis. Furthermore, by dynamically adjusting and logically updating the isolation response analysis set, and based on the changes in line load capacity and power supply path under the new topology, it reconstructs the protection coordination relationship and generates a power grid protection strategy migration mechanism. This enables adaptive switching of protection logic after fault isolation and improves stability and recovery efficiency under abnormal operating conditions.

[0117] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A fault handling method based on primary and secondary network coordination, characterized in that: include, Collect and preprocess power grid operation status data to generate a power grid operation data stream; Based on the power grid operation data stream, a graph neural network is used to perform topology identification and fault location of the power grid, and generate fault impact features. Based on the spatial range and electrical connection relationship of the fault impact characteristics, obtain the switches at both ends of the fault section of the power grid, perform fault isolation, and generate fault isolation information; The fault isolation information is analyzed using topology impact analysis. Combined with the changes in the current power grid topology, a power grid protection strategy migration mechanism is generated and a fault recovery process is initiated to restore power supply.

2. The fault handling method based on primary and distribution network coordination as described in claim 1, characterized in that: The power grid operation status data includes voltage parameters, current parameters, power parameters, switch status information, and protection device operation information; The preprocessing includes filtering and denoising, handling missing values, correcting outliers, and data standardization.

3. The fault handling method based on primary and distribution network coordination as described in claim 2, characterized in that: The generation of power grid operation data stream refers to the integration of preprocessed power grid operation status data using a multi-source heterogeneous data fusion method.

4. The fault handling method based on primary and distribution network coordination as described in claim 1, characterized in that: The specific steps for generating the power grid diagram are as follows: Extract the electrical parameters of each node in the power grid operation data stream, identify the equipment types in the power grid, and generate graph structure nodes; The topology connection data of the power grid is parsed using graph modeling to generate graph structure edges; By integrating the nodes and edges of the graph structure, a power grid graph is generated.

5. The fault handling method based on primary and distribution network coordination as described in claim 4, characterized in that: The specific steps for using graph neural networks to perform topology identification and fault location in the power grid are as follows: The graph neural network is used to embed the power grid graph and combine it with multi-channel time series information in the power grid operation data stream to generate node potential state vectors. Based on the node potential state vector, the regional division and line-transformer relationship in the power grid are identified, and the power grid topology is output by combining the electrical parameters in the power grid operation data stream. The connection relationships in the power grid topology are analyzed to generate power grid fault sections.

6. The fault handling method based on primary and distribution network coordination as described in claim 5, characterized in that: The generation of fault impact features refers to generating fault impact features by integrating the power grid topology and the faulty sections of the power grid.

7. The fault handling method based on primary and distribution network coordination as described in claim 1, characterized in that: The specific steps for obtaining the switches at both ends of the faulty section of the power grid based on the spatial range of the fault's impact characteristics and the electrical connection relationships are as follows: Extract the edge set from the power grid diagram and generate electrical connection relationships; The topology path tracing method is used to analyze the propagation path of the spatial range of fault impact characteristics, and adjacency matching is performed in combination with electrical connection relationships to generate the fault impact area. Identify the boundary points between the fault-affected area and the normal operating area, and obtain the boundary point switches through the switch status evaluation method to generate a switch set.

8. The fault handling method based on primary and distribution network coordination as described in claim 7, characterized in that: The specific steps for performing fault isolation and generating fault isolation information are as follows: The switch set is tripped, a remote tripping command is generated, and a safety verification is performed. The automated master station executes electrical isolation operations on remote tripping commands to generate fault isolation information.

9. The fault handling method based on primary and distribution network coordination as described in claim 1, characterized in that: The method of using topology impact analysis to analyze fault isolation information, combined with changes in the current power grid topology, generates a power grid protection strategy migration mechanism and initiates a fault recovery process to restore power supply. The specific steps are as follows: Based on fault isolation information, the changes in the current topology of the power grid are obtained, and the line load capacity in the fault-affected area is verified by combining power flow calculation method, and an isolation response analysis set is generated. Perform protection strategy generation and logic update operations on the isolation response analysis set to generate a power grid protection strategy migration mechanism; Based on the isolation response analysis set and the power grid protection strategy migration mechanism, a power supply restoration path is selected, and the fault recovery process is initiated.

10. A fault handling device based on primary and distribution network coordination, based on the fault handling method based on primary and distribution network coordination according to any one of claims 1 to 9, characterized in that: include, The data acquisition module is used to collect power grid operation status data and preprocess it to generate a power grid operation data stream; The identification and location module is used to perform topology identification and fault location of the power grid based on the power grid operation data stream and using a graph neural network to generate fault impact features. The fault isolation module is used to obtain the switches at both ends of the fault section of the power grid based on the spatial range and electrical connection relationship of the fault impact characteristics, and to perform fault isolation by remote control tripping method, generating fault isolation information; The startup analysis module is used to analyze fault isolation information using topology impact analysis. Combined with changes in the current power grid topology, it generates a power grid protection strategy migration mechanism and initiates a fault recovery process to restore power supply.

Images