A power transmission line fault knowledge graph reasoning method and system

Abstract

The application discloses a power transmission line fault knowledge graph reasoning method and system, relates to the technical field of knowledge graph reasoning, obtains the power transmission line fault knowledge graph stored in a detection area from a database, analyzes the power transmission line fault knowledge graph to obtain an initial power transmission line fault knowledge graph, judges whether the fault mode of the power transmission line is a new fault mode, extracts each cause-effect relationship under the fault mode if the fault mode is a new fault mode, regards each cause-effect relationship that does not exist in the initial power transmission line fault knowledge graph as each marked cause-effect relationship, determines the adding position of each marked cause-effect relationship in the initial power transmission line fault knowledge graph, verifies, updates the initial power transmission line fault knowledge graph, improves the reasoning efficiency, guarantees the correctness of the existing power transmission line fault knowledge graph, guarantees the effectiveness and accuracy of the adding position reasoning, and guarantees the accuracy of the fault reasoning result.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph reasoning technology, specifically to a knowledge graph reasoning method and system for power transmission line faults. Background Technology

[0002] Knowledge graph technology, with its powerful knowledge representation and reasoning capabilities, has been widely used in the field of transmission line fault diagnosis. By constructing a knowledge graph of transmission line faults, it is possible to quickly locate faults and analyze their causes, significantly improving the level of intelligent operation and maintenance of transmission lines. However, with the development of power systems towards high voltage and large capacity, as well as the application of energy grid integration and new materials, the fault modes faced by transmission lines are becoming increasingly diversified and complex, and new fault modes are constantly emerging. Therefore, there is an urgent need for a method and system for updating the knowledge graph of transmission line faults.

[0003] Traditional transmission line fault knowledge graph reasoning methods and systems rely on manual experience to determine the placement of causal relationships within the existing knowledge graph when a new fault mode occurs in a transmission line. Clearly, this approach has the following shortcomings: 1. Before determining the placement of causal relationships within the existing knowledge graph under a new fault mode, traditional methods and systems lack analysis of the existing transmission line fault knowledge graph, failing to ensure its accuracy. Consequently, the effectiveness of the reasoning for placement cannot be guaranteed in subsequent judgment processes.

[0004] 2. Traditional knowledge graph reasoning methods for transmission line faults rely on human experience to reason about the addition of causal relationships in new fault modes. This is not only inefficient but also prone to misjudgment and cannot guarantee the accuracy of reasoning at the added locations. When a transmission line fault occurs, the accuracy of the fault reasoning results cannot be guaranteed. Summary of the Invention

[0005] To address the aforementioned technical shortcomings, the present invention aims to provide a method and system for reasoning about knowledge graphs of transmission line faults.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: In the first aspect, the present invention provides a method for reasoning about a knowledge graph of transmission line faults, including the following steps: S1, graph acquisition: acquiring the transmission line fault knowledge graph already stored in the detection area from the database, analyzing it to remove erroneous information, and obtaining an initial transmission line fault knowledge graph.

[0007] S2. Data Acquisition and Analysis: Real-time monitoring of transmission line information within the detection area, and determination of whether the transmission lines in the detection area are faulty. If a fault is found, the fault mode of the transmission line is acquired, and analysis is performed to determine whether each fault mode is a new fault mode.

[0008] S3. Reasoning Knowledge Graph: When the fault mode is a new fault mode, extract the causal relationships under this fault mode, and refer to the causal relationships that do not exist in the initial transmission line fault knowledge graph as the labeled causal relationships. At the same time, based on the connection relationship between the nodes in the initial transmission line fault knowledge graph, determine the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph, and verify it.

[0009] S4. Update the knowledge graph: Update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results.

[0010] Secondly, the present invention provides a transmission line fault knowledge graph reasoning system, including the following modules: a graph acquisition module is used to acquire the transmission line fault knowledge graph already stored in the detection area from the database, analyze it to remove erroneous information, and acquire an initial transmission line fault knowledge graph.

[0011] The data acquisition and analysis module is used to monitor the information of transmission lines in the detection area in real time and determine whether the transmission lines in the detection area are faulty. If a fault is found, the fault mode of the transmission line is acquired and analyzed to determine whether each fault mode is a new fault mode.

[0012] The knowledge graph reasoning module is used to extract the causal relationships under a new fault mode when the fault mode is new. Each causal relationship that does not exist in the initial transmission line fault knowledge graph is called a labeled causal relationship. At the same time, based on the connection relationship between each node in the initial transmission line fault knowledge graph, the module determines the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph and verifies it.

[0013] The knowledge graph update module is used to update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results.

[0014] The database is used to store the transmission line fault knowledge graph, the mechanical-electrical relationship of the transmission line, the standard information characteristics of the transmission line, the characteristics of the head and tail nodes in each triplet, and the characteristics of each level in the transmission line fault knowledge graph that have been stored in the detection area.

[0015] The beneficial effects of this invention are as follows: 1. This invention provides a method and system for reasoning about a knowledge graph of transmission line faults. It retrieves the transmission line fault knowledge graph already stored in the detection area from the database, analyzes it to obtain an initial transmission line fault knowledge graph, and determines whether the fault mode of the transmission line is a new fault mode. If it is a new fault mode, it extracts the causal relationships under that fault mode, and refers to the causal relationships not present in the initial transmission line fault knowledge graph as labeled causal relationships. It determines the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph, verifies it, and updates the initial transmission line fault knowledge graph. This improves reasoning efficiency, ensures the correctness of the existing transmission line fault knowledge graph, guarantees the effectiveness and accuracy of reasoning at the addition position, and ensures the accuracy of the fault reasoning results.

[0016] 2. This invention refers to the existing knowledge graph of transmission line faults in the detection area in the database as the labeled transmission line fault knowledge graph. It obtains the mechanical-electrical relationship of the transmission line from the database, judges whether there are spurious correlations between the nodes in the labeled transmission line fault knowledge graph, and performs counterfactual simulation to obtain the initial transmission line fault knowledge graph. This ensures the correctness of the existing transmission line fault knowledge graph and the effectiveness of adding location reasoning.

[0017] 3. When the fault mode is a new fault mode, this invention obtains each triplet with a return value of 0 under this fault mode and refers to them as each labeled triplet. At the same time, based on the connection positions between each triplet and each node in the initial transmission line fault knowledge graph, the addition position of each labeled triplet in the initial transmission line fault knowledge graph is determined, and each labeled triplet is virtually added to the initial transmission line fault knowledge graph to obtain a virtual transmission line fault knowledge graph, which is then verified to ensure the accuracy of the reasoning for the addition position and the accuracy of the fault reasoning results. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. 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.

[0019] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention.

[0020] Figure 2 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 As shown, the present invention provides a method for reasoning about a knowledge graph of transmission line faults, including: S1, graph acquisition: acquiring the transmission line fault knowledge graph already stored in the detection area from the database, analyzing it to remove erroneous information, and obtaining an initial transmission line fault knowledge graph.

[0023] In a specific embodiment, the process of obtaining the initial transmission line fault knowledge graph is as follows: the transmission line fault knowledge graph of the detection area already stored in the database is called the labeled transmission line fault knowledge graph, and the mechanical-electrical relationship of the transmission line is obtained from the database. It is determined whether there is a pseudo-correlation between the nodes in the labeled transmission line fault knowledge graph. If there is a pseudo-correlation, the two nodes with pseudo-correlation are called pseudo-correlation node pairs. Each pseudo-correlation node pair in the labeled transmission line fault knowledge graph is obtained in this way, and each pseudo-correlation node pair in the labeled transmission line fault knowledge graph is removed to obtain the first transmission line fault knowledge graph.

[0024] It should be noted that each node in the transmission line fault knowledge graph represents environmental factors, fault types, and equipment levels.

[0025] It should also be noted that the physical constraints and parameter constraints between nodes are obtained from the knowledge graph of marked transmission line faults and compared with the mechanical-electrical relationship of the transmission line. If the physical constraints and parameter constraints between two nodes do not match the mechanical-electrical relationship of the transmission line, it means that the relationship between the two nodes is a pseudo-correlation. If the physical constraints and parameter constraints between two nodes match the mechanical-electrical relationship of the transmission line, it means that the relationship between the two nodes is not a pseudo-correlation. This method is used to determine whether there is a pseudo-correlation in the relationships between nodes in the knowledge graph of marked transmission line faults.

[0026] Suppose that there is a correlation between the high wind node and the conductor breakage node in the knowledge graph of transmission line faults, and the correlation is that the conductor breaks when the wind speed is 6 m / s. However, the mechanical-electrical relationship of the transmission line shows a correlation that the conductor breaks when the wind speed is 18 m / s. In this case, the correlation between the high wind node and the conductor breakage node in the knowledge graph of transmission line faults is a false correlation. This example is for illustrative purposes only and is not the only limitation.

[0027] Counterfactual simulations are performed on the knowledge graph of the first transmission line fault to verify the associations between the nodes in the knowledge graph. Associations between nodes that fail verification are removed to obtain the initial knowledge graph of the transmission line fault.

[0028] It should be noted that counterfactual simulation is an existing technology. The specific process is as follows: The causal relationship to be detected is extracted from the knowledge graph of the first transmission line fault. From this, the variables requiring intervention and the outcome variables are obtained. The variables requiring intervention are adjusted, and the outcome variables are observed to see if they change. If the outcome variables do not change, it means the causal relationship has not passed verification. If the outcome variables change, it is analyzed whether the change value of the outcome variables is greater than a threshold. If it is greater, it means the causal relationship has passed verification; if it is not less than a threshold, it means the causal relationship has not passed verification. It is important to know that counterfactual simulation is conducted in a simulated environment.

[0029] Assuming the causal relationship to be verified is that heavy rain causes short circuits in transmission lines, the variable that can be intervened is daily rainfall, and the outcome variable is insulator leakage current. The daily rainfall is reduced, and the insulator leakage current is observed. The change in insulator leakage current is calculated and compared to a threshold. If the change in insulator leakage current is greater than the threshold, the verification that heavy rain causes short circuits in transmission lines is passed; if the change in insulator leakage current is less than the threshold, the verification that heavy rain causes short circuits in transmission lines is not passed. This example is for illustrative purposes only and is not the only valid approach.

[0030] S2. Data Acquisition and Analysis: Real-time monitoring of transmission line information within the detection area, and determination of whether the transmission lines in the detection area are faulty. If a fault is found, the fault mode of the transmission line is acquired, and analysis is performed to determine whether each fault mode is a new fault mode.

[0031] It should be noted that the information of transmission lines includes current, voltage, power, phase, and surrounding environmental information, including wind speed, humidity, temperature, and precipitation.

[0032] It should also be noted that by collecting information such as current, voltage, power and phase through instruments such as fiber optic voltage transformers and electromagnetic transformers, this method and system are connected to a meteorological platform to obtain environmental information from the meteorological platform.

[0033] In a specific embodiment, the data acquisition and analysis process is as follows: real-time monitoring of the information of transmission lines in the detection area, acquisition of the information characteristics of transmission lines in the detection area, matching with the standard information characteristics in the database, determining whether the transmission lines in the detection area are faulty, and if the transmission lines in the detection area are faulty, acquisition of the fault cause and faulty facilities, and obtaining the fault mode of the transmission lines in the detection area.

[0034] It should be noted that the technique of obtaining information features of transmission lines within the detection area through LSTM is an existing technology. The specific process is as follows: the information of the transmission lines is filtered and missing values ​​are processed and normalized. The window size, sliding step size, input dimension and LSTM structure are set. The monitored data is divided into sample set, validation set and test set to train the LSTM. After training, the information of the transmission lines is input into the LSTM to obtain the information features of the transmission lines within the detection area.

[0035] It should also be noted that when the information characteristics of the transmission lines in the detection area are different from the standard information characteristics in the database, it indicates that the transmission lines in the detection area are faulty; when the information characteristics of the transmission lines in the detection area are the same as the standard information characteristics in the database, it indicates that the transmission lines in the detection area are not faulty.

[0036] Extract the causal relationships of the transmission lines within the detection area under this fault mode, compare them with the initial transmission line fault knowledge graph, and analyze whether the fault mode of the transmission lines within the detection area is a new fault mode.

[0037] The above-mentioned process for analyzing whether the fault mode of the transmission line in the detection area is a new fault mode is as follows: extract each causal relationship from the initial transmission line fault knowledge graph, and compile each causal relationship in the initial transmission line fault knowledge graph and each causal relationship of the transmission line in the detection area under the fault mode into triples, thereby obtaining each triple of the initial transmission line fault knowledge graph and each triple of the fault mode.

[0038] It should be noted that, assuming a causal relationship is that the conductor breaks when the wind speed reaches 18 m / s, then the triple of this causal relationship is (wind speed, conductor breakage, 18 m / s), where wind speed is the head node, conductor breakage is the tail node, and wind speed 18 m / s is the connecting relationship. This example is for illustrative purposes only and is not the only valid one.

[0039] Each triplet in the fault mode is compared with each triplet in the initial transmission line fault knowledge graph. The return value of each triplet in the fault mode is obtained. If there is a triplet with a return value of 0 in the fault mode, it means that the fault mode of the transmission line in the detection area is a new fault mode. If the return value of each triplet in the fault mode is 1, it means that the fault mode of the transmission line in the detection area is not a new fault mode.

[0040] It should be noted that each triplet in this fault mode is compared with each triplet in the initial transmission line fault knowledge graph. If a triplet in this fault mode is different from all triplets in the initial transmission line fault knowledge graph, the return value of that triplet in this fault mode is 0. If a triplet in this fault mode is the same as a triplet in the initial transmission line fault knowledge graph, the return value of that triplet in this fault mode is 1. The return values ​​of each triplet in this fault mode are obtained in this way.

[0041] In this context, two identical triples are defined as having the same head node, the same tail node, and the same connection relationship. Conversely, if they are not identical, the two triples are not identical. For example, if there are three triples: the first is (wind speed, conductor break, 18 m / s), the second is (wind speed, new composite conductor break, 18 m / s), and the third is (wind speed, conductor break, 18 m / s), then the first and third triples are identical, the first and second triples are not identical, and the third and second triples are also not identical. This example is for illustrative purposes only and is not the only valid definition.

[0042] S3. Reasoning Knowledge Graph: When the fault mode is a new fault mode, extract the causal relationships under this fault mode, and refer to the causal relationships that do not exist in the initial transmission line fault knowledge graph as the labeled causal relationships. At the same time, based on the connection relationship between the nodes in the initial transmission line fault knowledge graph, determine the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph, and verify it.

[0043] In a specific embodiment, the reasoning knowledge graph is implemented as follows: when the fault mode is a new fault mode, the triples that return a value of 0 under this fault mode are obtained and referred to as each labeled triple. At the same time, based on the connection positions between each triple and each node in the initial transmission line fault knowledge graph, the addition position of each labeled triple in the initial transmission line fault knowledge graph is determined.

[0044] Each labeled triplet is then virtually added to the initial transmission line fault knowledge graph to obtain a virtual transmission line fault knowledge graph. This graph is then verified, and the rationality of adding each labeled triplet is analyzed. If the addition is not rational, each labeled triplet that is not rationally added is called an adjustment triplet, and each adjustment triplet is stored separately.

[0045] It should be noted that the method for verifying the virtual power transmission knowledge graph is the same as the method for counterfactual simulation.

[0046] It should also be noted that if a labeled triple passes the validation, it means that the addition of the labeled triple is reasonable; if a labeled triple fails the validation, it means that the addition of the labeled triple is unreasonable.

[0047] The specific process for determining the addition position of each labeled triplet in the initial transmission line fault knowledge graph is as follows: extract the head node, tail node, and connection relationship of each labeled triplet, and compare the head node and tail node of each labeled triplet with each node in the initial transmission line fault knowledge graph.

[0048] When the head node or tail node of a certain labeled triple exists in the initial transmission line fault knowledge graph, the head node and tail node of the labeled triple are matched with each node in the initial transmission line fault knowledge graph to determine the position where the labeled triple is added.

[0049] When neither the head node nor the tail node of a labeled triple is in the initial transmission line fault knowledge graph, the characteristics of the head node and the tail node of the labeled triple are retrieved from the database and matched with the characteristics of each level in the transmission line fault knowledge graph in the database to determine the level of the labeled triple. A temporary connection is then made with the parent node of that level to determine the addition position of the labeled triple. This method is used to determine the addition position of each labeled triple in the initial transmission line fault knowledge graph.

[0050] It should be noted that the knowledge graph of transmission line faults includes levels such as equipment, environment, fault, and strategy. The characteristics of each level differ; for example, the characteristics of the equipment level include materials, model, and function, while the characteristics of the environment level are meteorological parameters. This example is for illustrative purposes only and is not the only valid approach.

[0051] It should also be noted that when the head and tail features of a labeled triple are the same as the features of a certain level, then that level is the level in which the labeled triple resides.

[0052] In the above, when the head node or tail node of a certain labeled triple exists in the initial transmission line fault knowledge graph, the addition position of the labeled triple is determined as follows: If both the head node and tail node of a certain labeled triple are in the initial transmission line fault knowledge graph, it means that the head node and tail node of the labeled triple exist in the initial transmission line. At this time, a connecting line is added between the head node and tail node of the labeled triple in the initial transmission line fault knowledge graph, and the link where the connecting line is located is analyzed. Then, the link is the addition position of the labeled triple.

[0053] It should be noted that the arrow of the connecting line points from the head node to the tail node.

[0054] If only one of the head and tail labels in a certain labeled triplet exists in the initial transmission line fault knowledge graph, then the node that exists in the initial transmission line fault knowledge graph is called a labeled node, and the other node is called a new node. A branch line is drawn from the labeled node to connect to the new node, and the link where the new node is located is analyzed according to the method of analyzing the link where the connection line is located. Then, this link is the location where the labeled triplet is added.

[0055] It should be noted that when the marked node is the head node, the previous node connected to the head node is retrieved.

[0056] The analysis of the link where the connection line is located, as described above, is carried out as follows: the physical constraints and parameter constraints between the head node and the tail node are obtained from the initial transmission line fault knowledge graph, as well as the physical constraints and parameter constraints between each node in the initial transmission line fault knowledge graph. Based on the physical constraints between the head node and the tail node, each node in the initial transmission line fault knowledge graph is traversed starting from the head node to obtain the initial link.

[0057] It should be noted that nodes whose physical constraints are the same as those between the head node and the tail node are selected for traversal.

[0058] In the initial link, each next node adjacent to the tail node is obtained and referred to as an associated node. At the same time, the state data of the tail node is obtained, and it is determined whether the state data of the tail node matches the parameter constraints between the tail node and each associated node. If the parameter constraints between the tail node and an associated node match the state data of the tail node, the associated node is called a marked associated node. If the parameter constraints between the tail node and an associated node do not match the state data of the tail node, the associated node is called an irrelevant node. Each marked associated node and each irrelevant node is obtained in this way.

[0059] It should be noted that the state data of the tail node refers to the parameter constraints between the tail node and its next node.

[0060] Remove all irrelevant nodes from the initial link, retain all marked nodes, and obtain the final link. This final link is the link to which the connection line is located.

[0061] In the above, when neither the head node nor the tail node of a certain labeled triple is in the initial transmission line fault knowledge graph, the addition position of the labeled triple is determined. The specific process is as follows: the level where the labeled triple is located is called the labeling level, and the physical constraints between the head node and the tail node in the labeled triple and the physical constraints between each node in the transmission line fault knowledge graph in the labeling level are obtained. The addition area of ​​the labeled triple is determined and called the labeling area.

[0062] It should be noted that the nodes whose physical constraints are the same as those between the head node and the tail node are obtained, and these nodes are clustered to obtain the region to be added for the labeled triple.

[0063] Within the marked area, temporary nodes will be created and connected to the parent node of the marked hierarchy. Constraint labels will be set on the connection lines, and it will be analyzed whether the constraint labels conflict with the constraint labels in the initial transmission line fault knowledge graph. If there is a conflict, the marked triplet will be stored separately.

[0064] It should be noted that the created temporary node is the labeled triple, and the constraint labels on the connection line are the connection relationships within the labeled triple. For example, suppose the constraint label on the connection line states that the wind speed reaches 18 m / s in a normal -45°C environment, while a constraint label in the initial transmission line fault knowledge graph states that the equipment malfunctions in a -20°C environment. In this case, the two labels conflict. This example is for illustrative purposes only and is not the only valid one.

[0065] S4. Update the knowledge graph: Update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results.

[0066] Please see Figure 2 As shown, the present invention provides a transmission line fault knowledge graph reasoning system, including: a graph acquisition module for acquiring the transmission line fault knowledge graph already stored in the detection area from the database, analyzing it to remove erroneous information, and acquiring an initial transmission line fault knowledge graph.

[0067] The data acquisition and analysis module is used to monitor the information of transmission lines in the detection area in real time and determine whether the transmission lines in the detection area are faulty. If a fault is found, the fault mode of the transmission line is acquired and analyzed to determine whether each fault mode is a new fault mode.

[0068] The knowledge graph reasoning module is used to extract the causal relationships under a new fault mode when the fault mode is new. Each causal relationship that does not exist in the initial transmission line fault knowledge graph is called a labeled causal relationship. At the same time, based on the connection relationship between each node in the initial transmission line fault knowledge graph, the module determines the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph and verifies it.

[0069] The knowledge graph update module is used to update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results.

[0070] The database is used to store the transmission line fault knowledge graph, the mechanical-electrical relationship of the transmission line, the standard information characteristics of the transmission line, the characteristics of the head and tail nodes in each triplet, and the characteristics of each level in the transmission line fault knowledge graph that have been stored in the detection area.

[0071] This invention retrieves the existing transmission line fault knowledge graph stored in the detection area from the database, analyzes it to obtain the initial transmission line fault knowledge graph, and determines whether the fault mode of the transmission line is a new fault mode. If it is a new fault mode, it extracts the causal relationships under this fault mode, and refers to the causal relationships that are not present in the initial transmission line fault knowledge graph as labeled causal relationships. It determines the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph, verifies it, and updates the initial transmission line fault knowledge graph. This improves reasoning efficiency, ensures the correctness of the existing transmission line fault knowledge graph, and also ensures the effectiveness and accuracy of the reasoning at the addition position, as well as the accuracy of the fault reasoning results.

[0072] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0073] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A knowledge graph reasoning method for transmission line faults, characterized in that, Includes the following steps: S1. Knowledge Graph Acquisition: Retrieve the stored knowledge graph of transmission line faults in the detection area from the database, analyze it to remove erroneous information, and obtain the initial knowledge graph of transmission line faults. S2. Data Acquisition and Analysis: Real-time monitoring of information on transmission lines within the detection area, and determination of whether the transmission lines in the detection area are faulty. If faulty, the fault mode of the transmission line is acquired, and each fault mode is analyzed to determine whether it is a new fault mode. S3. Reasoning Knowledge Graph: When the fault mode is a new fault mode, extract the causal relationships under this fault mode, and refer to the causal relationships that do not exist in the initial transmission line fault knowledge graph as the labeled causal relationships. At the same time, based on the connection relationships between the nodes in the initial transmission line fault knowledge graph, determine the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph, and verify it. The specific process of the reasoning knowledge graph is as follows: When the fault mode is a new fault mode, obtain the triples with a return value of 0 under this fault mode, and refer to them as the labeled triples. At the same time, based on the connection positions between each triple and each node in the initial transmission line fault knowledge graph, determine the addition position of each labeled triple in the initial transmission line fault knowledge graph. Each labeled triplet is then virtually added to the initial transmission line fault knowledge graph to obtain a virtual transmission line fault knowledge graph. This graph is then verified, and the rationality of adding each labeled triplet is analyzed. If the addition is not rational, each labeled triplet that is not rationally added is called an adjustment triplet, and each adjustment triplet is stored separately. The specific process for determining the addition position of each labeled triplet in the initial transmission line fault knowledge graph is as follows: extract the head node, tail node and connection relationship in each labeled triplet, and compare the head node and tail node in each labeled triplet with each node in the initial transmission line fault knowledge graph; When the head node or tail node of a certain labeled triple exists in the initial transmission line fault knowledge graph, the head node and tail node of the labeled triple are matched with each node in the initial transmission line fault knowledge graph to determine the position where the labeled triple is added. When neither the head node nor the tail node of a labeled triple is in the initial transmission line fault knowledge graph, the characteristics of the head node and the tail node of the labeled triple are obtained from the database and matched with the characteristics of each level in the transmission line fault knowledge graph in the database to determine the level of the labeled triple. A temporary connection is made with the parent node of the level to determine the addition position of the labeled triple. This method is used to determine the addition position of each labeled triple in the initial transmission line fault knowledge graph. S4. Update the knowledge graph: Update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results.

2. The method for reasoning about a knowledge graph of transmission line faults according to claim 1, characterized in that, The specific process for obtaining the initial transmission line fault knowledge graph is as follows: The existing knowledge graph of transmission line faults in the detection area in the database is called the labeled transmission line fault knowledge graph. The mechanical-electrical relationship of the transmission line is obtained from the database. It is determined whether there are pseudo-correspondences between the nodes in the labeled transmission line fault knowledge graph. If there are pseudo-correspondences, the two nodes with pseudo-correspondences are called pseudo-corresponding node pairs. Each pseudo-corresponding node pair in the labeled transmission line fault knowledge graph is obtained in this way. Each pseudo-corresponding node pair in the labeled transmission line fault knowledge graph is removed to obtain the first transmission line fault knowledge graph. Counterfactual simulations are performed on the knowledge graph of the first transmission line fault to verify the associations between the nodes in the knowledge graph. Associations between nodes that fail verification are removed to obtain the initial knowledge graph of the transmission line fault.

3. The method for reasoning about a knowledge graph of transmission line faults according to claim 1, characterized in that, The specific process of data acquisition and analysis is as follows: The system monitors the information of transmission lines within the detection area in real time, obtains the information characteristics of the transmission lines within the detection area, and matches them with the standard information characteristics in the database to determine whether the transmission lines in the detection area are faulty. If the transmission lines in the detection area are faulty, the system obtains the cause of the fault and the faulty facilities, and obtains the fault mode of the transmission lines within the detection area. Extract the causal relationships of the transmission lines within the detection area under this fault mode, compare them with the initial transmission line fault knowledge graph, and analyze whether the fault mode of the transmission lines within the detection area is a new fault mode.

4. The method for reasoning about a knowledge graph of transmission line faults according to claim 3, characterized in that, The specific process for analyzing whether the fault mode of the transmission lines within the detection area is a new fault mode is as follows: Extract each causal relationship from the initial transmission line fault knowledge graph, and compile each causal relationship in the initial transmission line fault knowledge graph and each causal relationship of the transmission line in the detection area under the fault mode into triplet, to obtain each triplet of the initial transmission line fault knowledge graph and each triplet of the fault mode. Each triplet in the fault mode is compared with each triplet in the initial transmission line fault knowledge graph. The return value of each triplet in the fault mode is obtained. If there is a triplet with a return value of 0 in the fault mode, it means that the fault mode of the transmission line in the detection area is a new fault mode. If the return value of each triplet in the fault mode is 1, it means that the fault mode of the transmission line in the detection area is not a new fault mode.

5. The method for reasoning about a knowledge graph of transmission line faults according to claim 1, characterized in that, When the head or tail node of a labeled triplet exists in the initial transmission line fault knowledge graph, the position for adding the labeled triplet is determined as follows: If both the head node and the tail node of a certain labeled triplet are in the initial transmission line fault knowledge graph, it means that the head node and the tail node of the labeled triplet exist in the initial transmission line. At this time, a connecting line is added between the head node and the tail node of the labeled triplet in the initial transmission line fault knowledge graph, and the link where the connecting line is located is analyzed. Then the link is the location where the labeled triplet is added. If only one of the head and tail labels in a certain labeled triplet exists in the initial transmission line fault knowledge graph, then the node that exists in the initial transmission line fault knowledge graph is called a labeled node, and the other node is called a new node. A branch line is drawn from the labeled node to connect to the new node, and the link where the new node is located is analyzed according to the method of analyzing the link where the connection line is located. Then, this link is the location where the labeled triplet is added.

6. The method for reasoning about a knowledge graph of transmission line faults according to claim 5, characterized in that, The specific process for analyzing the link containing this connection is as follows: The physical constraints and parameter constraints between the head node and the tail node, as well as the physical constraints and parameter constraints between each node in the initial transmission line fault knowledge graph, are obtained from the initial transmission line fault knowledge graph. Based on the physical constraints between the head node and the tail node, each node in the initial transmission line fault knowledge graph is traversed starting from the head node to obtain the initial link. In the initial link, each next node adjacent to the tail node is obtained and referred to as an associated node. At the same time, the state data of the tail node is obtained, and it is determined whether the state data of the tail node matches the parameter constraints between the tail node and each associated node. If the parameter constraints between the tail node and an associated node match the state data of the tail node, the associated node is called a marked associated node. If the parameter constraints between the tail node and an associated node do not match the state data of the tail node, the associated node is called an irrelevant node. Each marked associated node and each irrelevant node is obtained in this way. Remove all irrelevant nodes from the initial link, retain all marked nodes, and obtain the final link. This final link is the link to which the connection line is located.

7. The method for reasoning about a knowledge graph of transmission line faults according to claim 1, characterized in that, When both the head node and the tail node of a labeled triplet are not in the initial transmission line fault knowledge graph, the position for adding the labeled triplet is determined as follows: The level where the labeled triplet is located is called the labeling level. The physical constraints between the head node and the tail node in the labeled triplet and the physical constraints between each node in the transmission line fault knowledge graph in the labeling level are obtained. The region where the labeled triplet is added is determined and called the labeling region. Within the marked area, temporary nodes will be created and connected to the parent node of the marked hierarchy. Constraint labels will be set on the connection lines, and it will be analyzed whether the constraint labels conflict with the constraint labels in the initial transmission line fault knowledge graph. If there is a conflict, the marked triplet will be stored separately.

8. A transmission line fault knowledge graph reasoning system for implementing the transmission line fault knowledge graph reasoning method according to any one of claims 1-7, characterized in that, include: The knowledge graph acquisition module is used to retrieve the transmission line fault knowledge graph already stored in the detection area from the database, analyze it to remove erroneous information, and obtain the initial transmission line fault knowledge graph. The data acquisition and analysis module is used to monitor the information of the transmission lines in the detection area in real time and determine whether the transmission lines in the detection area are faulty. If there is a fault, the fault mode of the transmission line is acquired and analyzed to determine whether each fault mode is a new fault mode. The knowledge graph reasoning module is used to extract causal relationships under a new fault mode when the fault mode is new. Causal relationships not present in the initial transmission line fault knowledge graph are referred to as labeled causal relationships. Simultaneously, based on the connection relationships between nodes in the initial transmission line fault knowledge graph, the module determines the addition position of each labeled causal relationship in the initial transmission line fault knowledge graph and verifies it. The specific process of reasoning the knowledge graph is as follows: When the fault mode is new, the module obtains triples with a return value of 0 under this fault mode and refers to them as labeled triples. Simultaneously, based on the connection positions between each triple and node in the initial transmission line fault knowledge graph, the module determines the addition position of each labeled triple in the initial transmission line fault knowledge graph. Each labeled triplet is then virtually added to the initial transmission line fault knowledge graph to obtain a virtual transmission line fault knowledge graph. This graph is then verified, and the rationality of adding each labeled triplet is analyzed. If the addition is not rational, each labeled triplet that is not rationally added is called an adjustment triplet, and each adjustment triplet is stored separately. The specific process for determining the addition position of each labeled triplet in the initial transmission line fault knowledge graph is as follows: extract the head node, tail node and connection relationship in each labeled triplet, and compare the head node and tail node in each labeled triplet with each node in the initial transmission line fault knowledge graph; When the head node or tail node of a certain labeled triple exists in the initial transmission line fault knowledge graph, the head node and tail node of the labeled triple are matched with each node in the initial transmission line fault knowledge graph to determine the position where the labeled triple is added. When neither the head node nor the tail node of a labeled triple is in the initial transmission line fault knowledge graph, the characteristics of the head node and the tail node of the labeled triple are obtained from the database and matched with the characteristics of each level in the transmission line fault knowledge graph in the database to determine the level of the labeled triple. A temporary connection is made with the parent node of the level to determine the addition position of the labeled triple. This method is used to determine the addition position of each labeled triple in the initial transmission line fault knowledge graph. The knowledge graph update module is used to update the initial transmission line fault knowledge graph based on the addition position of each initial causal relationship in the initial transmission line fault knowledge graph and the verification results. The database is used to store the transmission line fault knowledge graph, the mechanical-electrical relationship of the transmission line, the standard information characteristics of the transmission line, the characteristics of the head and tail nodes in each triplet, and the characteristics of each level in the transmission line fault knowledge graph that have been stored in the detection area.

Images