Power communication network fault diagnosis method and system based on time-constrained graph matching

Abstract

This application discloses a method and system for diagnosing power communication network faults based on time-constrained graph matching, comprising: generating a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed; determining typical types of power communication network faults, and generating a power communication network fault query graph containing time-constrained relationships under different faults based on the topological structure of the power communication network to be diagnosed; storing candidate edges and their partial matching relationships in a preset index structure; searching the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraints of the power communication network fault query graph based on the preset index structure and using an incremental matching algorithm in the generated matching order; and determining the defect fault results of the power communication network to be diagnosed based on the power communication network data subgraph. This invention improves the efficiency and accuracy of power communication network fault diagnosis.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart grid fault diagnosis, and in particular relates to a method and system for diagnosing faults in a power communication network based on time-constrained graph matching. Background Art

[0002] Power communication networks are a vital component of modern power systems. Their primary function is to provide reliable communication support for the coordination, monitoring, and control of power equipment. With the rapid development of smart grids, power communication networks are playing an increasingly crucial role in power system operations. They must not only implement basic communication functions but also support complex tasks such as power dispatching, real-time monitoring, and protection and control. The realization of these functions relies on the high reliability, low latency, and high bandwidth performance of power communication networks. However, the complexity and dynamic nature of power communication networks also make them susceptible to various defects and failures.

[0003] In actual operation, power communication networks may face various problems, such as communication delays, data loss, device anomalies, and network topology changes. Communication delays can occur due to network congestion, link quality degradation, or device performance issues, impacting the real-time control and protection capabilities of the power system. Data loss can be caused by network interference or device anomalies, hindering the transmission of critical information. Device anomalies, such as failures in routers, switches, or fiber links, can cause network outages, impacting overall system operation. Furthermore, network topology changes due to device maintenance, fault isolation, or adjustments to load balancing strategies further complicate fault diagnosis. These issues not only reduce the operational efficiency of power communication networks but also pose a threat to the stability and security of the power system. For example, communication delays can prevent protection devices from operating in a timely manner, leading to cascading failures or even large-scale power outages. Therefore, rapidly and accurately diagnosing defects and faults in power communication networks is crucial for ensuring the safe operation of power systems.

[0004] Despite the importance of fault diagnosis in power communication networks, this process faces numerous challenges. First, power communication networks are typically multi-layered, multi-device, and multi-protocol, resulting in large scale and complex structures. Faults can occur on devices or links at different levels, increasing the difficulty of diagnosis. Second, the operating state of power communication networks is dynamic, and fault propagation is both real-time and random. Traditional static diagnostic methods struggle to adapt to this dynamic environment. Furthermore, some faults may be hidden, such as performance degradation on some links or intermittent equipment anomalies, making them difficult to detect directly through traditional monitoring methods. Finally, the amount of data generated in power communication networks is enormous and may contain a significant amount of noise. Extracting useful fault information from this massive amount of data is a crucial research issue.

[0005] To address these difficulties, researchers have proposed a variety of fault diagnosis methods, including rule-based diagnosis, data-driven machine learning methods, and graphical model-based analysis methods. However, these methods still have certain limitations when faced with complex power communication networks. For example, rule-based methods rely on expert experience and have difficulty adapting to complex dynamic environments; while data-driven methods can learn patterns from data, they may have difficulty explaining the root cause of faults in the absence of prior knowledge. Summary of the Invention

[0006] To address the deficiencies in the prior art, the present invention provides a method and system for power communication network fault diagnosis based on time-constrained graph matching. In a dynamic environment containing multiple communication devices and complex network topologies, the method and system can quickly and accurately identify the root cause and propagation path of the fault, thereby improving the efficiency and accuracy of power communication network fault diagnosis.

[0007] The present invention adopts the following technical solutions.

[0008] A first aspect of the present invention provides a method for diagnosing faults in a power communication network based on time-constrained graph matching, comprising:

[0009] S1: generating a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed;

[0010] S2: Determine typical types of power communication network faults, and generate a power communication network fault query graph containing time constraints under different faults based on the topology of the power communication network to be diagnosed;

[0011] S3: Based on the edge labels and time constraints of the power communication network fault query graph, all candidate edges are found from the power communication network data graph to be diagnosed, and partial matching relationships between the candidate edges are recorded. The candidate edges and their partial matching relationships are stored in a preset index structure. Based on the preset index structure, an incremental matching algorithm is used to search the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraints of the power communication network fault query graph in the generated matching order.

[0012] S4: Determine the defect and fault results of the power communication network to be diagnosed according to the power communication network data subgraph.

[0013] Optionally, in S3, the preset index structure includes node groups and links, each node group includes edges in the power communication network fault query graph and all candidate edges in the power communication network data graph to be diagnosed that match the edges in the power communication network fault query graph, and the links represent the connection relationship between the candidate edges in the power communication network data graph to be diagnosed, and the connection relationship satisfies the structural constraints and time sequence constraints in the power communication network fault query graph.

[0014] Optionally, multiple candidate edges with the same neighbor structure are merged into one node in the preset index structure.

[0015] Optionally, in S3, when time-constrained matching is performed using an incremental matching algorithm, when an edge in the power communication network data graph to be diagnosed is updated, a corresponding update operation is performed in a preset index structure by searching for partial matches related to the edge update.

[0016] Optionally, in S3, the matching order generated includes:

[0017] Calculate the number of connected nodes for each node in the preset index structure;

[0018] Sort the partially matched relationships according to the number of connected nodes for each node;

[0019] Generate matching order according to the sorting results.

[0020] Optionally, in S3, searching the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraint of the power communication network fault query graph in the generated matching order by using an incremental matching algorithm based on a preset index structure includes:

[0021] Recursively searching for candidate edges that partially match each edge in the power communication network fault query graph in a preset index structure according to a matching order through an incremental matching algorithm, and adding the found candidate edges to a dynamic matching set;

[0022] If the dynamic matching set corresponding to the first candidate edge found does not contain all the edges in the power communication network fault query graph, then delete the first candidate edge found in the dynamic matching set, and backtrack to the second candidate edge that partially matches the corresponding edge in the power communication network fault query graph to continue time constraint matching;

[0023] If the dynamic matching set contains all the edges in the electric power communication network fault query graph, the recursion is stopped to obtain the electric power communication network data subgraph that matches the time constraint of the electric power communication network fault query graph.

[0024] Optionally, in S4, determining the defect and fault result of the power communication network to be diagnosed according to the power communication network data subgraph includes:

[0025] Calculating the matching degree between multiple power communication network data subgraphs and the query graph respectively;

[0026] Select the electric power communication network data subgraph corresponding to the maximum matching degree;

[0027] Defect and fault results of the power communication network to be diagnosed are determined from the selected power communication network data subgraph.

[0028] Optionally, respectively calculating the matching degree between multiple power communication network data subgraphs and the query graph includes:

[0029] Calculate the structural matching degree and temporal matching degree of multiple power communication network data subgraphs and the query graph respectively;

[0030] The structural matching degree and the temporal matching degree are weightedly summed to obtain the matching degree between multiple power communication network data subgraphs and the query graph.

[0031] A second aspect of the present invention provides a power communication network fault diagnosis system, the system comprising:

[0032] A first generating module is configured to generate a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed;

[0033] The second generation module is used to determine the typical type of power communication network fault and generate a power communication network fault query graph containing time constraints under different faults based on the topology of the power communication network to be diagnosed;

[0034] A time-constrained graph matching module is used to find all candidate edges from the power communication network data graph to be diagnosed based on the edge labels and time constraints of the power communication network fault query graph, record the partial matching relationships between the candidate edges, store the candidate edges and their partial matching relationships in a preset index structure, and query the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraints of the power communication network fault query graph in the generated matching order using an incremental matching algorithm based on the preset index structure;

[0035] The determination module is used to determine the defect and fault results of the power communication network to be diagnosed based on the power communication network data subgraph.

[0036] The third aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is loaded into the processor, the method for diagnosing faults in a power communication network based on time-constrained graph matching is implemented.

[0037] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned power communication network fault diagnosis method based on time-constrained graph matching.

[0038] Compared with the prior art, the beneficial effects of the present invention include at least:

[0039] This paper proposes a fault diagnosis method for power communication networks based on time-constrained graph matching. By modeling the fault propagation process as a dynamic sequential process and combining network topology with a graph matching algorithm, it achieves accurate identification of the fault root cause and propagation path. Unlike traditional deep learning methods, this method does not rely on large amounts of manually annotated data, thus avoiding the time-consuming and labor-intensive manual labeling. At the same time, by introducing time constraints, it fully utilizes the temporal characteristics of fault propagation and defines a series of temporal constraint relationships, further improving the efficiency and accuracy of diagnosis. In dynamic environments containing multiple communication devices and complex network topologies, the combination of time constraints and graph matching algorithms demonstrates greater flexibility and adaptability, particularly for rapidly adapting to topological changes and real-time diagnostic requirements in power communication networks. Experimental results demonstrate the high feasibility and practicality of this method, demonstrating its significant value in promoting the intelligentization of power communication networks. It also reveals the potential for widespread application of time-constrained modeling and graph matching techniques in future power communication network management, significantly reducing the cost and complexity of maintenance and operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0041] Figure 1 This is a flow chart of a method for diagnosing faults in a power communication network based on time-constrained graph matching, provided by an embodiment of the present invention;

[0042] Figure 2 This is another flowchart of a method for diagnosing faults in a power communication network based on time-constrained graph matching provided by an embodiment of the present invention;

[0043] Figure 3 This is another flowchart of a method for diagnosing faults in a power communication network based on time-constrained graph matching provided by an embodiment of the present invention;

[0044] Figure 4This is a schematic diagram of an application of a method for diagnosing faults in a power communication network based on time-constrained graph matching provided by an embodiment of the present invention;

[0045] Figure 5 This is a typical network topology diagram of a power communication network provided by an embodiment of the present invention;

[0046] Figure 6 This is a schematic diagram of a query graph topology provided by an embodiment of the present invention;

[0047] Figure 7 This is a schematic diagram of query matching speed comparison provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0048] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0049] The following are definitions related to the graph matching problem in an embodiment of the present invention.

[0050] Definition 1 Graph: A directed, connected, labeled graph is denoted by .in, is a set of vertices, is the edge set, is a collection of tags, is a labeling function that labels each vertex or edge Map to A label in or .for The two edges in , if they share a common endpoint, they are said to and are adjacent. For each edge , its neighbor set is recorded as ,Right now The degree of the edge Defined as the number of its neighbors.

[0051] In the present invention, directed, connected, labeled graphs are collectively referred to as graphs.

[0052] Definition 2 Subgraph Isomorphism: Given a query graph and a data graph If there is a arrive Injective mapping of m , such that:

[0053] 1. For any ,have ; 2. For any ,have and , then it is called yes The subgraphs of .

[0054] Will be from Vertex in to The injective mapping of the vertices in is called exist Isomorphic embedding of subgraphs in . Query graph and data graph The two edges in and , if satisfied ,and and (or and ), then it is called and is a matching edge, denoted as .

[0055] Definition 3 Flow Graph: Flow Graph It is a growing graph consisting of a series of newly added edges, denoted as ,in Indicates that from the vertex To the top A directed edge with label and arrival time For two edges and ,if , then it is called Before , recorded as To simplify the representation, this paper only focuses on edge insertion operations in the stream graph. It should be noted that our method can be easily extended to support edge deletion operations based on the time sliding window model.

[0056] Definition 4 Time Order: Given a query graph , its time sequence It is defined in A strict partial order relation on . For two edges ,if , then it means in the data graph and The matching edges must also satisfy the partial order relation on arrival time.

[0057] if Not in O O In the definition, it is called is disordered, denoted as To simplify the presentation, this article will be presented in chronological order. O The query graph of .

[0058] Definition 5: Time Constraint Matching: Given a query graph and data graph A subgraph in ,if If the following two conditions are met, then yes A time-constrained match of:

[0059] Structural constraints (isomorphism): yes exist A structure embedded in .

[0060] Time order constraints: For Any two edges in ,if , then in The corresponding edge The time sequence must also meet .

[0061] Definition 6 Partial matching: refers to the The query graph found in Partially similar substructures. A partial match result is a To the data graph A partial mapping that satisfies the following two conditions:

[0062] Structural constraints (partial isomorphism): The edges and vertices in the partial matching results are structurally identical to the query graph. That is, the edges and vertices in the partial matching results are isomorphic to the data graph. A query graph is formed in A subgraph with the same structure as part of .

[0063] Temporal order constraint: The edges in the partial matching results not only match in structure but also satisfy the query graph The time order constraint defined in . If the query graph Edge Prior to the edge in time , then in the partial matching results, Matched edges It must also precede in time Matched edges .

[0064] Combine Figure 1 and Figure 4 As shown, embodiment 1 of the present invention provides a method for diagnosing faults in a power communication network based on time-constrained graph matching, which specifically includes the following contents:

[0065] S1: Based on the topological structure of the power communication network to be diagnosed, a data graph of the power communication network to be diagnosed is generated.

[0066] The data graph of the power communication network to be diagnosed can be expressed as ,in, represents the vertex set of the power communication network data graph to be diagnosed, represents the edge set of the power communication network data graph to be diagnosed, Indicates mapping vertices or edges to The label function, Represents a set of labels, including vertex type, edge type, and edge time information.

[0067] The time information of an edge can be a time window condition, such as the alarm trigger time range, alarm arrival time, etc.

[0068] Specifically, raw data such as optical routing tables, defect ticket tables, and defect ticket alarm tables in the power communication network are obtained. A data dictionary is used to define standardized fields (such as device type and port number) to ensure data consistency. Using a non-relational graph database, the data is abstracted into nodes and relationships to construct a comprehensive data graph of the communication network. Using sites, devices, and ports as nodes and optical path connections as edges, the topology of the power communication network to be diagnosed is constructed by mapping optical paths with site, device, and port information. Defect alarm events are then mapped as edges in the graph, each with a timestamp and label.

[0069] Combine Figure 5 As shown, V ( G )={Substation A, Substation B, Substation C, Equipment A, Equipment B, Equipment C, Equipment D, A(1), A(2), B(1), B(2), C(1), C(2), C(3), C(4), C(5)}, E ( G)={(Substation A, Equipment A),(Equipment A, Equipment A(1)),(Equipment A, Equipment A(2)),(Substation A, Substation B),(Substation B, Equipment B),(Equipment B, Equipment B(1)),(Equipment B, Equipment B(2)),(Substation B, Substation C),(Substation C, Equipment C),(Equipment C, Equipment C(1)),(Equipment C, Equipment C(2)),(Equipment C, Equipment C(3)),(Substation C, Equipment D),(Equipment D, Equipment C(4)),(Equipment D, Equipment C(5)),(A(1),B(1)),(B(2),C(1)),(A(2),C(5))}. The equipment in the system and its connection methods are represented by a set of nodes and edges.

[0070] S2: Determine typical types of power communication network faults, and generate a power communication network fault query graph containing time constraints under different faults based on the topology of the power communication network to be diagnosed.

[0071] The fault query graph of the power communication network can be expressed as ,in, represents the vertex set of the power communication network fault query graph, Represents the edge set of the power communication network fault query graph.

[0072] Specifically, by analyzing the defect and fault information, communication network faults are obtained, and the communication network defect and faults include transmission optical path interruption, optical cable interruption, equipment power supply interruption, equipment board fault, and equipment power supply fault.

[0073] It is understandable that these faults correspond to different power communication network fault query diagrams, and fault categories are queried through different features.

[0074] Temporal constraints refer to restrictions imposed on the temporal attributes of edges or nodes in a data graph during the matching process. They can be used to describe the temporal patterns of fault propagation. These temporal attributes can include edge or node creation time, update time, or expiration time. For example, upper and lower bounds on propagation delay can be defined to represent the time required for a fault to propagate from one node to another.

[0075] Combine Figure 6 As shown, Figure 6 The corresponding query graph is shown in Table 1.

[0076] Table 1

[0077]

[0078] In Table 1, the first module queries the basic information of edges in the graph. This information consists of (edge ID, start vertex ID, target vertex ID, edge attribute label, start vertex attribute label, target vertex attribute label). For example, the first row (e 0 0 3 10 1) indicates that the edge is present. The 0 represents the unique identifier of the edge, used to distinguish different edges. The second 0 represents the ID of the start vertex of the edge, and the 3 represents the ID of the target vertex of the edge. The first 1 describes the edge's attribute or type, indicating the type of edge. In this case, it is a device connection. In the initial modeling, edge types were simplified to three: optical cable connection, device connection, and port connection, corresponding to labels 0, 1, and 2, respectively. The third 0 represents the label of the start vertex. There are three vertex types: site, device, and port, corresponding to labels 0, 1, and 2, respectively. The final 0 and 1 represent the labels of the start and target vertex, respectively. The second module adds time constraint information. The b here indicates that this is a time constraint. Taking (b 3 4 2 1 0) as an example, the edge with ID 3 must arrive before 4, 4 arrives before 2, 2 arrives before 1, and 1 arrives before 0. This series of edge IDs indicates that these edges must occur in the specified order.

[0079] Combine Figure 2 As shown, in S2, the typical type of power communication network fault is determined, and based on the topology of the power communication network to be diagnosed, a power communication network fault query diagram under different faults is generated, including:

[0080] Obtain historical defect alarm information;

[0081] Analyze historical defect alarm information and screen out nodes and links affected by faults, determine the coverage of defect alarm information and typical types of power communication network faults;

[0082] In the topology of the power communication network to be diagnosed, nodes within the coverage of the defect alarm information are connected to generate a power communication network fault query graph, and time information is added to the edges in the query graph.

[0083] Historical defect alarm information includes defect order alarm information and defect order information. Defect order alarm information records the alarm details when the fault is triggered, and defect order information stores the fault reporting content, such as time information, location information, and fault alarm type information.

[0084] In this embodiment, first, by analyzing the timing characteristics of faults, the fault search scope can be quickly narrowed, thereby improving diagnostic efficiency. Second, time-constrained modeling can combine network topology and time information to accurately identify the root cause and propagation path of the fault. Furthermore, this method can adapt to the dynamic changes of power communication networks and is suitable for real-time fault diagnosis scenarios. Finally, the diagnostic results based on time constraints are highly interpretable, helping operations and maintenance personnel better understand the mechanism and propagation of faults.

[0085] S3: According to the edge labels and time constraint relationships of the power communication network fault query graph, all candidate edges are found from the power communication network data graph to be diagnosed, and the partial matching relationships between the candidate edges are recorded. The candidate edges and their partial matching relationships are stored in a preset index structure. Based on the preset index structure, an incremental matching algorithm is used to query the power communication network data subgraph that matches the time constraint of the power communication network fault query graph in the power communication network data graph to be diagnosed in the generated matching order.

[0086] Optionally, the preset index structure includes node groups and links, each node group includes edges in the power communication network fault query graph and all candidate edges in the power communication network data graph to be diagnosed that match the edges in the power communication network fault query graph, and the links represent the connection relationship between the candidate edges in the power communication network data graph to be diagnosed, and the connection relationship satisfies the structural constraints and time sequence constraints in the power communication network fault query graph.

[0087] Specifically, a pre-defined index structure groups nodes. Each group corresponds to an edge in the query graph and includes all edges in the data graph that match that edge. Links are created between nodes to encapsulate partial matches (i.e., partial embeddings) and temporal ordering information of edges in the data graph. If two nodes from different groups share a common endpoint and adhere to temporal constraints, a corresponding link is added between them.

[0088] It can be understood that when nodes in different groups are adjacent, and their corresponding mapping edges are also adjacent, and the arrival time of the nodes complies with the time order defined by the mapping edge, a corresponding link is added between the two nodes.

[0089] Specifically, you can initialize a node set with a preset index structure and edge sets Then, process the data graph in ascending order of arrival time. to construct the edge and Specifically, given the data graph An edge in Matching edges in the query graph ,examine Is it an unordered edge, or for Each predecessor edge Does it exist? The corresponding node in . is not empty, the partial order must be established, and the edges can be processed in order. Zhongwei Create a node, then Create corresponding links between nodes. Specifically, consider exist Each adjacent edge in , and retrieve its matching edges .if and The edge relationship and time order between them meet the conditions (nodes in different groups have common endpoints and follow the time constraint information), then find The middle corresponds to the edge Nodes, and and After building the preset index structure, set The status of the midpoint.

[0090] The partial matching results are stored by constructing a preset index structure, where the partial matching results include structural constraint and time constraint information of the query graph.

[0091] Specifically, building a preset index structure can be completed in the offline stage and can be completed through a one-time operation, which reduces the backtracking search space and memory consumption, thereby improving the efficiency of locating candidate matches in the online stage.

[0092] Optional, combined Figure 3 As shown, multiple candidate edges with the same neighbor structure are merged into one node in the preset index structure.

[0093] Specifically, during the traversal process, if the adjacency relationships of multiple candidate edges are completely consistent in the current path, they are merged. This can reduce the number of candidates, eliminate redundant candidate paths, reduce redundant calculations, and reduce computational complexity, thereby improving the efficiency of incremental matching.

[0094] Optionally, a node in the preset index structure includes a node status attribute, and the node status attribute is used to indicate whether all adjacent edges of the node in the data graph have been matched.

[0095] Specifically, in the online stage, based on the preset index structure, an incremental matching algorithm is used to search the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraint of the power communication network fault query graph in accordance with the generated matching order.

[0096] Specifically, the pre-set index structure is dynamically updated online. First, the query edge corresponding to the newly inserted edge is determined. Its timestamp is verified to meet the temporal order constraints, for example, whether it is later than the matching edge of the predecessor edge. If so, the newly inserted edge is added to the corresponding group and links are established with neighboring edges. If the new link causes the node to meet all neighbor conditions, its node status attribute is updated to 1.

[0097] In this way, during the matching phase, only the nodes whose matching node status attribute is 1 in the preset index structure are traversed, where 1 indicates a match. This can be dynamically updated to ensure that the preset index structure reflects the data graph status in real time, and only the layout is adjusted to avoid reconstruction overhead, thereby reducing redundant partial matches.

[0098] Optionally, when the timestamp of an edge of the data graph of the power communication network to be diagnosed exceeds a preset time sliding window range, the edge of the data graph is deleted from the data graph of the power communication network to be diagnosed.

[0099] Specifically, initialize the time sliding window size , preset time sliding window Represents a time range, indicating the time range of the data currently considered. If the time sliding window size Set to 10 minutes, Indicates the data within 10 minutes before the current time. Record each edge in the data graph , record its timestamp, which indicates the occurrence time, update time or arrival time of the edge, and calculate the starting time of the preset time sliding window as the current time minus , traverse each edge of the data graph. If its timestamp is less than the start time of the preset time sliding window, delete the corresponding edge and update the data graph.

[0100] Optionally, in S3, when time-constrained matching is performed using an incremental matching algorithm, when an edge in the power communication network data graph to be diagnosed is updated, a corresponding update operation is performed in a preset index structure by searching for partial matches related to the edge update.

[0101] Specifically, update operations include edge insertion and edge deletion. When processing an edge insertion operation, the newly inserted edge is first added to the data graph, and the preset index structure is updated to include relevant information about the new edge. Subsequently, all incremental matching results containing the new edge are searched based on the preset index structure, ensuring that these matches meet the structural constraints and temporal order constraints of the query graph. When processing an edge deletion operation, all incremental matching results containing the edge to be deleted are first searched and recorded. Subsequently, the edge to be deleted is removed from the data graph, and the preset index structure is updated to remove the candidate edges and partial matching relationships related to the edge.

[0102] In this way, each time the data graph is updated, only the partial matches related to the updated edges are processed through incremental calculation. The relevant candidate edges and partial matches are quickly located through the preset index structure, avoiding the recalculation of the entire graph and improving the efficiency of time-constrained matching. In addition, all incremental matching results of the query graph can be output in real time.

[0103] Optionally, the generated matching order includes:

[0104] Calculate the number of connected nodes for each node in the preset index structure;

[0105] Sort the partially matched relationships according to the number of connected nodes for each node;

[0106] Generate matching order according to the sorting results.

[0107] Specifically, connected nodes refer to other nodes directly connected to nodes in the preset index structure, represented by links. Sorting partial matches based on the number of connected nodes for each node means sorting partial matches by the number of connected nodes they correspond to, placing edges with fewer connected nodes first. Based on the sorted results, edges from the query graph corresponding to the nodes are added to the matching order. This prioritizes partial matches with fewer connected nodes, allowing for faster identification of feasible and infeasible matches.

[0108] Specifically, we can start from the node corresponding to the newly inserted edge as the trigger point for incremental update, and give priority to the query edge with the least number of candidate edges in the current partial match for expansion to reduce branches.

[0109] You can also use the edge adjacency and time order constraints stored in the preset index structure to extend the matching along the index links. If the candidate edges of a path are pruned, the algorithm will automatically skip the branch and generate a new order according to the remaining links.

[0110] Optionally, the generated matching order includes:

[0111] The matching order is determined by combining the frequency of occurrence of each node in the data graph and the number of connected nodes of each node in the preset index structure.

[0112] Optionally, the matching order is determined by combining the frequency of occurrence of each node in the data graph and the number of connected nodes of each node in the preset index structure, including:

[0113] Calculate the priority index based on the frequency of each node in the data graph and the number of connected nodes of each node in the preset index structure;

[0114] The matching order is determined according to the priority index.

[0115] Specifically, the priority index is calculated according to the frequency of each node in the data graph and the number of connected nodes of each node in the preset index structure:

[0116] ,

[0117] in, It is a priority index. The larger the priority index value is, the higher the priority is and the earlier it is matched. It is the normalized value of the frequency of nodes in the preset index structure appearing in the data graph. The normalized value of the number of connected nodes of the node in the preset index structure is used to match nodes with high frequency in the data graph and small number of connected nodes. 、 are the corresponding weights, is a small constant.

[0118] It is understandable that the matching order determines the order in which edges are expanded during the matching process, which directly affects the size of the search space and matching efficiency. During the matching search, we start with a partial match with fewer connected nodes and gradually expand the partial match until a complete match is found or it is determined that it cannot be expanded. This way, by prioritizing partial matches with fewer connected nodes rather than considering all edges in the data graph, we reduce the search space and improve matching efficiency. Furthermore, the nodes and links in the preset index structure already encapsulate the edges and adjacency relationships in the data graph, allowing for more efficient incremental matching. Furthermore, by dynamically selecting the starting point and book order, we ensure that the search starts from the path that is most likely to form a complete match, improving efficiency.

[0119] Optionally, in S3, searching the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraint of the power communication network fault query graph in the generated matching order by using an incremental matching algorithm based on a preset index structure includes:

[0120] Recursively searching for candidate edges that partially match each edge in the power communication network fault query graph in a preset index structure according to a matching order through an incremental matching algorithm, and adding the found candidate edges to a dynamic matching set;

[0121] If the dynamic matching set corresponding to the first candidate edge found does not contain all the edges in the power communication network fault query graph, then delete the first candidate edge found in the dynamic matching set, and backtrack to the second candidate edge that partially matches the corresponding edge in the power communication network fault query graph to continue time constraint matching;

[0122] If the dynamic matching set contains all the edges in the electric power communication network fault query graph, the recursion is stopped to obtain the electric power communication network data subgraph that matches the time constraint of the electric power communication network fault query graph.

[0123] Specifically, each edge in the query graph is processed step by step in the matching order. For each edge in the query graph, all possible candidate edges are found from the preset index structure. These candidate edges are the edges that match the data graph and the query graph. These candidate edges are added to the partial matching result, and then the next edge is expanded according to the matching order until the partial matching result includes all edges in the query graph or cannot be expanded.

[0124] It can be understood that the dynamic matching set is a dynamically constructed temporary data structure used to record the current matching status.

[0125] S4: Determine the defect and fault results of the power communication network to be diagnosed according to the power communication network data subgraph.

[0126] S4 includes:

[0127] Calculating the matching degree between multiple power communication network data subgraphs and the query graph respectively;

[0128] Select the electric power communication network data subgraph corresponding to the maximum matching degree;

[0129] Defect and fault results of the power communication network to be diagnosed are determined from the selected power communication network data subgraph.

[0130] Specifically, the matching degree of multiple power communication network data subgraphs is analyzed, and the power communication network data subgraph that covers the most alarm events and meets the time constraint is preferentially selected to determine the defect and fault results of the power communication network to be diagnosed. The defect and fault results of the power communication network to be diagnosed include faulty devices or links.

[0131] Optionally, respectively calculating the matching degree between multiple power communication network data subgraphs and the query graph includes:

[0132] Calculate the structural matching degree and temporal matching degree of multiple power communication network data subgraphs and the query graph respectively;

[0133] The structural matching degree and the temporal matching degree are weightedly summed to obtain the matching degree between multiple power communication network data subgraphs and the query graph.

[0134] Optionally, the matching degree between multiple power communication network data subgraphs and the query graph is calculated as follows:

[0135] ,

[0136] in, is the i-th power communication network data subgraph Power communication network fault query diagram The matching degree, is the weight of the structural matching degree, ranging from 0 to 1. is the structural matching degree between the i-th power communication network data subgraph and the query graph, for The alarm events in The degree of temporal matching of the edges or vertices in the network is selected. The power communication network data subgraph with the highest matching degree is selected to analyze and determine the fault results of the power communication network to be diagnosed.

[0137] Optionally, the structural matching degree between the i-th electric power communication network data subgraph and the query graph is calculated as follows:

[0138] ,

[0139] in, Query diagram for power communication network fault The edge set of Power communication network data subgraph The edge set of Query diagram for power communication network fault and power communication network data subgraph The number of common edges.

[0140] Optional, calculated as follows The alarm events in The degree of temporal matching of the edges or vertices in :

[0141] ,

[0142] in, For alarm events timestamp, for Zhongyu The timestamp of the associated edge or vertex, is the indicator function, if If yes, it is 1, otherwise it is 0. Is a set of alarm events, each alarm event is associated with is associated with an edge or vertex in .

[0143] S4 specifically includes determining the fault type and corresponding quantity based on the power communication network data subgraph obtained by graph matching, and providing the range of defect information.

[0144] In power communication systems, fault phenomena can generally be categorized into three types. The first type involves faults limited to a single network element. These faults do not affect other network elements and are typically detected by the faulty element and its network management system. In this case, at least two nodes are affected. The second type involves a single network element fault that propagates to other network elements, resulting in a wider impact. The third type involves a failure in the communication medium between network elements, such as a break in an optical cable or optical path, which typically affects multiple nodes or sites. For both types of faults, the farthest node affected is typically the boundary of the fault's impact range—that is, the farthest network element reporting the fault. Time constraints also play a key role in diagnosing both types of faults. By analyzing the temporal sequence of alarms, the fault's propagation path can be determined, and the fault's starting point and scope can be further inferred, enabling precise location and diagnosis of complex faults.

[0145] It should be noted that if the queried power communication network data subgraph is a subgraph with only one vertex and / or a subgraph with more than two vertices and a node degree of 1, the power communication network does not have any defects or faults. This can be used as a filter condition for pruning during matching, reducing invalid matches and improving fault diagnosis efficiency.

[0146] In this embodiment, the judgment method for device board failure and device power supply failure in communication network defect faults is relatively special. The device board failure characteristics are that they are issued by the same board or different boards under the same device, while the device power supply failure is issued by the board. Generally, a substation board issues POWER ABNORMAL or a substation protection network issues POWER FAIL. Their failures are all board-level failures, and the failure point is relatively single. Therefore, these two types of easily identifiable faults can be directly queried or the method provided by the embodiment of the present invention can be used for fault query. The main difference between transmission optical path interruption and optical cable interruption is at the substation site level. If there is no connection between substations, it is an optical cable interruption, while if there is no connection between substations, it is an optical path interruption.

[0147] Therefore, based on the above situation in the application context of power communication network defect diagnosis, the embodiment of the present invention excludes some special cases. For data graph subgraphs that meet any of the following conditions, the embodiment of the present invention does not regard these as cases where defect faults exist: a subgraph with only one vertex, no connecting edges connected to it, a subgraph with more than 2 vertices and a node degree of 1, and no alarm subgraph.

[0148] Embodiment 2 of the present invention provides a power communication network fault diagnosis system, which runs the power communication network fault diagnosis method based on time-constrained graph matching as described in embodiment 1. The system includes:

[0149] A first generating module is configured to generate a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed;

[0150] The second generation module is used to determine the typical type of power communication network fault and generate a power communication network fault query graph containing time constraints under different faults based on the topology of the power communication network to be diagnosed;

[0151] A time-constrained graph matching module is used to find all candidate edges from the power communication network data graph to be diagnosed based on the edge labels and time constraints of the power communication network fault query graph, record the partial matching relationships between the candidate edges, store the candidate edges and their partial matching relationships in a preset index structure, and query the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraints of the power communication network fault query graph in the generated matching order using an incremental matching algorithm based on the preset index structure;

[0152] The determination module is used to determine the defect and fault results of the power communication network to be diagnosed based on the power communication network data subgraph.

[0153] Embodiment 3 of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded into the processor, the method for diagnosing faults in a power communication network based on time-constrained graph matching described in embodiment 1 is implemented.

[0154] Embodiment 4 of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for diagnosing faults in a power communication network based on time-constrained graph matching according to embodiment 1 is implemented.

[0155] Example 5 of the present invention provides a verification example of the method described in Example 1.

[0156] Combine Figure 7 As shown in the figure, we analyze a large-scale local power communication network in China. This network has 208 nodes and 297 computational edges. The algorithm below is run on an Intel(R) Core i9-13900HX CPU with a 5.40GHz clock speed, 16GB of memory, and an Ubuntu 18.04 operating system.

[0157] To investigate the application effectiveness of graph matching technology in target communication networks, this paper compares and analyzes the performance differences between graph matching technology and existing relational databases used in power grids when processing complex, connected data. In relational databases, a common SQL operation is to combine the first rows of two or more tables through a join. This approach is very useful when processing complex, connected data. However, as data scale increases, especially when multiple tables are joined, join operations can become time-consuming due to performance bottlenecks, making efficient processing difficult. In contrast, for large-scale communication networks, graph matching technology offers superior scalability and performance advantages when processing complex topological relationships and multi-dimensional data. In particular, in communication networks, fault alarm data often exhibits strong correlations and a certain temporal order. Graph matching technology can more efficiently capture these complex relationships and further optimize the matching process through time constraints, thereby improving the accuracy and efficiency of fault diagnosis.

[0158] Duplicate defect ticketing often occurs in actual defect ticket data, a phenomenon that needs to be reduced and optimized through data processing. Duplicate defect tickets primarily arise from two situations: First, in real-world environments, due to defects not being detected and fixed promptly, the same issue triggers an alarm again some time later. Due to the long interval between alarms, traditional systems often treat multiple alarms caused by the same root cause as separate independent events, resulting in multiple defect reports. By introducing time constraints, alarm timestamps can be analyzed to determine whether these alarms are duplicate events within the same time range. For example, if the interval between two alarms is less than a certain threshold, they can be considered the same event, effectively reducing the generation of duplicate defect tickets. Second, when multiple geographically dispersed sites share the same line or equipment, if a fault occurs on that line or equipment, the diagnostic system, based on manual judgment or preset rules, may mistakenly interpret the fault as multiple independent defects, generating multiple independent defect records. By introducing time constraints, we can analyze whether the alarms at these sites occur in a concentrated manner within a short period of time, based on the chronological order of alarm occurrence. This can help determine whether these alarms originate from the same fault cause and allow the relevant defect tickets to be merged.

[0159] The introduction of time constraints plays a significant role in resolving the issue of duplicate defect tickets. By comparing timestamps and analyzing time intervals, alarm events with the same root cause can be grouped together, avoiding misjudgments caused by long time intervals or overly dispersed alarm distribution. Furthermore, combined with the topological analysis capabilities of graph matching technology, correlations between alarms can be more efficiently identified. Especially in complex communication networks, the combination of time constraints and graph matching technology can more comprehensively capture the propagation path and impact range of alarms, providing a more accurate basis for defect diagnosis and optimized dispatch.

[0160] Experimental results show that the response time of traditional relational databases is much slower than that of graph matching algorithms when faced with a variety of different fault types. Among the more novel graph matching algorithms, the graph matching algorithm in Example 1 is more efficient in time processing than the rapidflow algorithm. Among the three fault type queries, the power supply terminal is more efficient in time querying for transmission optical path interruption and optical cable interruption than the other two fault types because its characteristics are more distinct than the graph matching patterns of the other two types. Experimental comparisons show that graph matching is more effective for fault diagnosis.

[0161] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0162] The present disclosure may be a system, method and / or computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0163] A computer-readable storage medium can be a tangible device that can hold and store instructions for use by an instruction execution device. A computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punched card or raised structure in a groove on which instructions are stored, and any suitable combination thereof. As used herein, a computer-readable storage medium is not to be construed as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse passing through a fiber optic cable), or an electrical signal transmitted through an electrical wire.

[0164] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0165] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, the state information of the computer-readable program instructions is used to personalize an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), so that the electronic circuit can execute the computer-readable program instructions, thereby implementing various aspects of the present disclosure.

[0166] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for fault diagnosis of power communication network based on time-constrained graph matching, characterized in that: include: S1: generating a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed; S2: Determine typical types of power communication network faults, and based on the topology of the power communication network to be diagnosed, generate a power communication network fault query graph containing time constraints under different faults; wherein the time constraints are used to describe the time pattern of fault propagation; S3: Based on the edge labels and time constraints of the power communication network fault query graph, all candidate edges are found from the power communication network data graph to be diagnosed, and partial matching relationships between the candidate edges are recorded. The candidate edges and their partial matching relationships are stored in a preset index structure. Based on the preset index structure, an incremental matching algorithm is used to search the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraints of the power communication network fault query graph in the generated matching order. The preset index structure includes node groups and links, each node group includes edges in the power communication network fault query graph and all candidate edges in the power communication network data graph to be diagnosed that match the edges in the power communication network fault query graph, and the links represent the connection relationship between the candidate edges in the power communication network data graph to be diagnosed, and the connection relationship satisfies the structural constraints and time order constraints in the power communication network fault query graph; S4: Determine the defect and fault results of the power communication network to be diagnosed according to the power communication network data subgraph.

2. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 1, characterized in that: Multiple candidate edges with the same neighbor structure are merged into one node in a preset index structure.

3. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 1, characterized in that: In S3, when time-constrained matching is performed using the incremental matching algorithm, when an edge in the power communication network data graph to be diagnosed is updated, a corresponding update operation is performed in the preset index structure by searching for partial matches related to the edge update.

4. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 1 or 2, characterized in that: In S3, the matching order is generated as follows: Calculate the number of connected nodes for each node in the preset index structure; Sort the partially matched relationships according to the number of connected nodes for each node; Generate matching order according to the sorting results.

5. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 4, characterized in that: In S3, based on the preset index structure, an incremental matching algorithm is used to search the power communication network data graph to be diagnosed for a power communication network data subgraph that matches the time constraint of the power communication network fault query graph in accordance with the generated matching order, including: Recursively searching for candidate edges that partially match each edge in the power communication network fault query graph in a preset index structure according to a matching order through an incremental matching algorithm, and adding the found candidate edges to a dynamic matching set; If the dynamic matching set corresponding to the first candidate edge found does not contain all the edges in the power communication network fault query graph, then delete the first candidate edge found in the dynamic matching set, and backtrack to the second candidate edge that partially matches the corresponding edge in the power communication network fault query graph to continue time constraint matching; If the dynamic matching set contains all the edges in the electric power communication network fault query graph, the recursion is stopped to obtain the electric power communication network data subgraph that matches the time constraint of the electric power communication network fault query graph.

6. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 1, characterized in that: In S4, the defect and fault results of the power communication network to be diagnosed are determined according to the power communication network data subgraph, including: Calculating the matching degree between multiple power communication network data subgraphs and the query graph respectively; Select the electric power communication network data subgraph corresponding to the maximum matching degree; Defect and fault results of the power communication network to be diagnosed are determined from the selected power communication network data subgraph.

7. The method for fault diagnosis of a power communication network based on time-constrained graph matching according to claim 6, characterized in that: Calculate the matching degree between multiple power communication network data subgraphs and the query graph respectively, including: Calculate the structural matching degree and temporal matching degree of multiple power communication network data subgraphs and the query graph respectively; The structural matching degree and the temporal matching degree are weightedly summed to obtain the matching degree between multiple power communication network data subgraphs and the query graph.

8. A power communication network fault diagnosis system, characterized in that: The system comprises: A first generating module is configured to generate a data graph of the power communication network to be diagnosed based on the topological structure of the power communication network to be diagnosed; The second generation module is configured to determine typical types of power communication network faults and, based on the topology of the power communication network to be diagnosed, generate a power communication network fault query graph containing time constraints for different faults; wherein the time constraints are used to describe the temporal pattern of fault propagation; A time constraint graph matching module is used to find all candidate edges from the power communication network data graph to be diagnosed based on the edge labels and time constraint relationships of the power communication network fault query graph, record the partial matching relationships between the candidate edges, and store the candidate edges and their partial matching relationships in a preset index structure. Based on the preset index structure, an incremental matching algorithm is used to query the power communication network data subgraph that matches the time constraint of the power communication network fault query graph in the power communication network data graph to be diagnosed in accordance with the generated matching order; wherein the preset index structure includes node groups and links, each node group includes edges in the power communication network fault query graph and all candidate edges in the power communication network data graph to be diagnosed that match the edges in the power communication network fault query graph, and the links represent the connection relationships of the candidate edges in the power communication network data graph to be diagnosed, and the connection relationships satisfy the structural constraints and time sequence constraints in the power communication network fault query graph; The determination module is used to determine the defect and fault results of the power communication network to be diagnosed based on the power communication network data subgraph.

9. An electronic device comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the power communication network fault diagnosis method based on time-constrained graph matching according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the power communication network fault diagnosis method based on time-constrained graph matching described in any one of claims 1 to 7 are implemented.

Images