Operation and maintenance fault processing method and device based on large model and knowledge graph, equipment and medium

Abstract

The invention discloses an operation and maintenance fault processing method and device based on a large model and a knowledge graph, equipment and a medium, and relates to the field of operation and maintenance fault diagnosis processing, and the method comprises the steps: constructing an operation and maintenance fault processing domain ontology, and inputting historical operation and maintenance fault information and the operation and maintenance fault processing domain ontology into a preset large model to obtain a target triple, obtaining an operation and maintenance fault processing knowledge graph based on the target triple; performing dependency syntactic analysis on the historical operation and maintenance fault information to obtain a dependency syntactic tree, processing the dependency syntactic tree based on the operation and maintenance fault processing knowledge graph to obtain a first dependency syntactic tree, and obtaining a fault processing question and answer template library by using the first dependency syntactic tree; determining a second dependency syntax tree based on the operation and maintenance fault information to be processed, determining a question and answer template corresponding to the second dependency syntax tree from a question and answer template library, and instantiating the question and answer template to obtain a semantic query graph; and determining a target sub-graph corresponding to the semantic query graph from the knowledge graph, and obtaining an operation and maintenance fault processing result based on the target sub-graph.

Description

Technical Field

[0001] This application relates to the field of operation and maintenance fault diagnosis and processing, and particularly to an operation and maintenance fault processing method, device, equipment and medium based on a large model and a knowledge graph. Background Art

[0002] With the wide application and rapid development of information technology, the scale and complexity of various information systems are constantly increasing, and operation and maintenance management work has become increasingly heavy and complex. Traditional operation and maintenance fault diagnosis and processing methods mainly rely on the experience judgment of operation and maintenance personnel, which is not only inefficient but also error-prone. In addition, on the one hand, most existing operation and maintenance fault management systems store historical operation and maintenance fault diagnosis and processing cases in the form of databases or files, lacking the exploration of the potential value and systematic management of historical operation and maintenance fault diagnosis and processing cases, resulting in the waste of knowledge resources as valuable fault diagnosis and processing experience is difficult to be effectively integrated and reused. On the other hand, these systems generally lack intelligent fault diagnosis and processing mechanisms and cannot quickly identify the root cause of faults and automatically recommend accurate solutions, further increasing the operation and maintenance management cost. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide an operation and maintenance fault processing method, device, equipment and medium based on a large model and a knowledge graph, which can make full use of the advantages of the large model and the knowledge graph to solve the problems of imperfect management of historical fault diagnosis and processing cases, low efficiency and easy errors in current operation and maintenance work, so as to realize the automation, high efficiency and precision of operation and maintenance fault diagnosis and processing. The specific solutions are as follows:

[0004] In the first aspect, this application provides an operation and maintenance fault processing method based on a large model and a knowledge graph, including:

[0005] Constructing an operation and maintenance fault processing domain ontology based on historical operation and maintenance fault information, and inputting the historical operation and maintenance fault information and the operation and maintenance fault processing domain ontology into a preset large model to obtain target triples output by the preset large model, and obtaining an operation and maintenance fault processing knowledge graph based on the target triples; the operation and maintenance fault processing domain ontology is used to describe the types, properties of each data in the historical operation and maintenance fault information and the relationships between the data;

[0006] Performing dependency syntax analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntax tree, and performing standardization processing on the dependency syntax tree based on the operation and maintenance fault processing knowledge graph to obtain a first dependency syntax tree, and obtaining a fault processing question and answer template library by using the first dependency syntax tree; the fault processing question and answer template library includes a target dependency syntax tree after removing entity information from the first dependency syntax tree and a corresponding question and answer template;

[0007] Determine a second dependency syntax tree based on the operation and maintenance fault information to be processed sent by the user terminal, and use the target dependency syntax tree to determine a question-and-answer template corresponding to the second dependency syntax tree from the fault handling question-and-answer template library, and instantiate the question-and-answer template according to the operation and maintenance fault information to be processed to obtain a semantic query graph;

[0008] Determine a target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and obtain an operation and maintenance fault handling result based on the target subgraph.

[0009] Optionally, constructing an operation and maintenance fault handling domain ontology based on historical operation and maintenance fault information includes:

[0010] Extract concept object information, data attribute information, and object attribute information from historical operation and maintenance fault information, and construct an operation and maintenance fault handling domain ontology based on the concept object information, the data attribute information, and the object attribute information;

[0011] Among them, the concept object information includes alarm events, alarm names, and alarm solutions; the data attribute information includes alarm times and IP addresses; the object attribute information includes first attribute information representing the occurrence of an alarm, second attribute information representing the association between different alarms, and third attribute information representing the alarm handling method.

[0012] Optionally, inputting the historical operation and maintenance fault information and the operation and maintenance fault handling domain ontology into a preset large model to obtain a target triple output by the preset large model includes:

[0013] Assign entity category labels to the concept object information and the data attribute information, assign relationship category labels to the object attribute information, and input the concept object information, the data attribute information, the object attribute information, the corresponding labels, and the historical operation and maintenance fault information into the preset large model;

[0014] Perform entity recognition operations and relationship extraction operations based on the preset large model to extract operation and maintenance fault information defined by the operation and maintenance fault handling domain ontology from the historical operation and maintenance fault information, and output it in the form of a triple to obtain a target triple.

[0015] Optionally, standardizing the dependency syntax tree based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntax tree includes:

[0016] For the dependency syntax tree, use entity matching technology based on similarity calculation to determine a matching subgraph with the highest similarity to the dependency syntax tree from the operation and maintenance fault handling knowledge graph;

[0017] Complete entity alignment between the dependency syntax tree and the matching subgraph based on rule extraction technology, and use the dependency syntax tree after entity alignment as the first dependency syntax tree;

[0018] Correspondingly, obtaining the fault handling Q&A template library using the first dependency syntax tree includes:

[0019] Perform entity information removal operation on the first dependency syntax tree to obtain a target dependency syntax tree, and determine Q&A templates based on concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntax tree;

[0020] Determine the fault handling Q&A template library based on the target dependency syntax tree and the corresponding Q&A templates.

[0021] Optionally, determining a second dependency syntax tree based on the to-be-processed operation and maintenance fault information sent by the user terminal, and using the target dependency syntax tree to determine the Q&A template corresponding to the second dependency syntax tree from the fault handling Q&A template library includes:

[0022] Obtain the to-be-processed operation and maintenance fault information sent by the user terminal, determine the second dependency syntax tree corresponding to the to-be-processed operation and maintenance fault information, and perform entity information removal on the second dependency syntax tree to obtain a second target dependency syntax tree;

[0023] Determine the target dependency syntax tree corresponding to the second target dependency syntax tree from the fault handling Q&A template library to obtain the corresponding Q&A template.

[0024] Optionally, instantiating the Q&A template according to the to-be-processed operation and maintenance fault information to obtain a semantic query graph includes:

[0025] Determine the target template in the Q&A template that is not represented by the target wildcards, and instantiate the target template based on the to-be-processed operation and maintenance fault information to obtain the semantic query graph corresponding to the Q&A template.

[0026] Optionally, determining the target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and obtaining the operation and maintenance fault handling result based on the target subgraph includes:

[0027] Determine the target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, determine the similarity between the semantic query graph and each target subgraph, and select the target subgraph with the highest similarity as the final target subgraph;

[0028] Use the alarm solution corresponding to the final target subgraph as the operation and maintenance fault handling result corresponding to the to-be-processed operation and maintenance fault information.

[0029] In a second aspect, the present application provides an operation and maintenance fault handling device based on a large model and a knowledge graph, including:

[0030] A knowledge graph determination module, configured to construct an ontology for operation and maintenance fault handling based on historical operation and maintenance fault information, and input the historical operation and maintenance fault information and the ontology for operation and maintenance fault handling into a preset large model to obtain target triples output by the preset large model, and obtain an operation and maintenance fault handling knowledge graph based on the target triples; the ontology for operation and maintenance fault handling is used to describe the types, properties of each data in the historical operation and maintenance fault information, and the relationships between the data;

[0031] A template library determination module, configured to perform dependency syntactic analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntactic tree, perform standardization processing on the dependency syntactic tree based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntactic tree, and use the first dependency syntactic tree to obtain a fault handling question and answer template library; the fault handling question and answer template library includes a target dependency syntactic tree after removing entity information from the first dependency syntactic tree and a corresponding question and answer template;

[0032] A semantic query graph determination module, configured to determine a second dependency syntactic tree based on the operation and maintenance fault information to be processed sent by the user side, use the target dependency syntactic tree to determine a question and answer template corresponding to the second dependency syntactic tree from the fault handling question and answer template library, and instantiate the question and answer template according to the operation and maintenance fault information to be processed to obtain a semantic query graph;

[0033] A result determination module, configured to determine a target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and obtain an operation and maintenance fault handling result based on the target subgraph.

[0034] In a third aspect, the present application provides an electronic device, including:

[0035] A memory, configured to store a computer program;

[0036] A processor, configured to execute the computer program to implement the foregoing operation and maintenance fault handling method based on a large model and a knowledge graph.

[0037] In a fourth aspect, the present application provides a computer-readable storage medium, configured to store a computer program, wherein the computer program, when executed by a processor, implements the foregoing operation and maintenance fault handling method based on a large model and a knowledge graph.

[0038] In this application, an ontology in the field of operation and maintenance fault handling is constructed based on historical operation and maintenance fault information, and the historical operation and maintenance fault information and the ontology in the field of operation and maintenance fault handling are input into a preset large model to obtain target triples output by the preset large model, and an operation and maintenance fault handling knowledge graph is obtained based on the target triples; the ontology in the field of operation and maintenance fault handling is used to describe the types, properties of each data in the historical operation and maintenance fault information, and the relationships between the data; dependency syntactic analysis is performed on the historical operation and maintenance fault information to obtain a corresponding dependency syntactic tree, and the dependency syntactic tree is standardized based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntactic tree, and the first dependency syntactic tree is used to obtain a fault handling question and answer template library; the fault handling question and answer template library includes a target dependency syntactic tree after removing entity information from the first dependency syntactic tree and a corresponding question and answer template; a second dependency syntactic tree is determined based on the operation and maintenance fault information to be processed sent by the user terminal, and the question and answer template corresponding to the second dependency syntactic tree is determined from the fault handling question and answer template library by using the target dependency syntactic tree, and the question and answer template is instantiated according to the operation and maintenance fault information to be processed to obtain a semantic query graph; a target subgraph corresponding to the semantic query graph is determined from the operation and maintenance fault handling knowledge graph, and an operation and maintenance fault handling result is obtained based on the target subgraph. As can be seen from the above, in this application, an ontology in the field of operation and maintenance fault handling is constructed, and an operation and maintenance fault handling knowledge graph is obtained based on the ontology in the field of operation and maintenance fault handling and a preset large model, which is used for the effective management and efficient reuse of historical operation and maintenance fault information; the dependency syntactic tree is standardized based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntactic tree, and the first dependency syntactic tree is used to obtain a fault handling question and answer template library, so that based on the fault handling question and answer template library, for the operation and maintenance fault information to be processed, it can be quickly mapped to the corresponding question and answer template to generate a semantic query graph, and through the subgraph matching technology, a target subgraph corresponding to the semantic query graph is determined from the operation and maintenance fault handling knowledge graph, and an operation and maintenance fault handling result is obtained based on the target subgraph, so as to realize the instant response and efficient processing of operation and maintenance faults, be able to identify the root cause of faults more quickly and accurately, and at the same time reduce the dependence on manual experience in operation and maintenance fault diagnosis work and improve the operation and maintenance work efficiency. Description of the Drawings

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained according to the provided drawings without creative efforts.

[0040] Figure 1 It is a flowchart of an operation and maintenance fault handling method based on a large model and a knowledge graph disclosed in the present application;

[0041] Figure 2 Schematic diagram for generating a fault handling Q&A template disclosed in the present application;

[0042] Figure 3 Schematic diagram for handling operation and maintenance fault problems disclosed in the present application;

[0043] Figure 4 Overall schematic diagram of an operation and maintenance fault handling method based on a large model and a knowledge graph disclosed in the present application;

[0044] Figure 5 Schematic diagram of the structure of an operation and maintenance fault handling device based on a large model and a knowledge graph disclosed in the present application;

[0045] Figure 6 Schematic diagram of the structure of an electronic device disclosed in the present application. Detailed implementation manners

[0046] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0047] With the wide application and rapid development of information technology, the scale and complexity of various information systems have been continuously increasing, and operation and maintenance management work has become increasingly heavy and complex. Traditional operation and maintenance fault diagnosis and handling methods mainly rely on the experience judgment of operation and maintenance personnel, which is not only inefficient but also error-prone. In addition, on the one hand, most of the existing operation and maintenance fault management systems store historical operation and maintenance fault diagnosis and handling cases in the form of databases or files, lacking the mining of the potential value and systematic management of historical operation and maintenance fault diagnosis and handling cases, resulting in the valuable fault diagnosis and handling experience being difficult to be effectively integrated and reused, causing a waste of knowledge resources. On the other hand, these systems generally lack an intelligent fault diagnosis and handling mechanism and cannot quickly identify the root cause of the fault and automatically recommend accurate solutions, further increasing the operation and maintenance management cost. Therefore, the present application provides an operation and maintenance fault handling method based on a large model and a knowledge graph, which can make full use of the advantages of the large model and the knowledge graph to solve the problems of imperfect management of historical fault diagnosis and handling cases, low efficiency and error-proneness in current operation and maintenance work, so as to realize the automation, high efficiency and precision of operation and maintenance fault diagnosis and handling.

[0048] Refer to Figure 1 As shown, the embodiments of the present application disclose an operation and maintenance fault handling method based on a large model and a knowledge graph, including:

[0049] Step S11: Construct an ontology for operation and maintenance fault handling based on historical operation and maintenance fault information, and input the historical operation and maintenance fault information and the ontology for operation and maintenance fault handling into a preset large model to obtain target triples output by the preset large model, and obtain an operation and maintenance fault handling knowledge graph based on the target triples; the ontology for operation and maintenance fault handling is used to describe the types, properties of each data in the historical operation and maintenance fault information, and the relationships between the data.

[0050] In this embodiment, first, historical operation and maintenance fault information is obtained. The data forms in the historical operation and maintenance fault information may include structured data and unstructured data, and the data content may include key alarm information such as alarm name, alarm area, alarm component, alarm level, and alarm solution. Then, an ontology for operation and maintenance fault handling is constructed based on the historical operation and maintenance fault information, which may include extracting concept object information, data attribute information, and object attribute information from the historical operation and maintenance fault information, and constructing an ontology for operation and maintenance fault handling based on the concept object information, data attribute information, and object attribute information; among them, the concept object information includes, but is not limited to, alarm events, alarm names, and alarm solutions. These concept objects are the core elements for constructing the domain ontology and represent various entities in the operation and maintenance fault handling field. For example, an alarm event is a broad concept that includes various specific fault alarm situations; the alarm name is a specific description of each alarm event; the alarm solution is the handling method for different alarm events; the data attribute information includes, but is not limited to, alarm time, IP address, etc. These data attributes assign specific values or information to the concept objects, making the concept objects more specific and recognizable. For example, the alarm time can be used to understand the sequence and time pattern of fault occurrences, and the IP address can be used to locate the network device where the fault occurs; the object attribute information includes, but is not limited to, the first attribute information representing the occurrence of an alarm, such as "occur", "trigger", "generate", etc., the second attribute information representing the association between different alarms, such as "associate", "cause", "accompany", etc., and the third attribute information representing the alarm handling method, such as "handle", "repair", "optimize", etc. By constructing an ontology for operation and maintenance fault handling, the historical operation and maintenance fault information can be organized in a structured and systematic manner, and the deep association value between the historical operation and maintenance fault information can be mined and revealed.

[0051] It should be noted that when constructing the ontology for operation and maintenance fault diagnosis and handling, the Protégé tool can be applied, and based on the "Stanford Seven-Step Method", the concept object information, data attribute information, and object attribute information can be extracted to construct the ontology for operation and maintenance fault diagnosis and handling.

[0052] After constructing the ontology in the field of operation and maintenance fault handling, the historical operation and maintenance fault information and the ontology in the field of operation and maintenance fault handling can be input into a preset large model to obtain the target triples output by the preset large model. Specifically, entity category labels are assigned to the conceptual object information and data attribute information, relationship category labels are assigned to the object attribute information, and the conceptual object information, data attribute information, object attribute information, the corresponding labels, and the historical operation and maintenance fault information are input into the preset large model; then, entity recognition operations and relationship extraction operations are performed based on the preset large model to extract the operation and maintenance fault information defined by the operation and maintenance fault handling ontology from the historical operation and maintenance fault information and output it in the form of triples to obtain the target triples.

[0053] It can be understood that the historical operation and maintenance fault information is complex and diverse. To enable the large model to better identify the historical operation and maintenance fault information, labels can be assigned to different types of data. For example, "server overheating" is marked as the "fault phenomenon" entity category, "October 1, 2024" is marked as the "time" entity category, "causes" is marked as the "causal relationship" relationship category, and "belongs to" is marked as the "subordinate relationship" relationship category. In this way, the large model can quickly identify the categories to which these information belong according to the labels. For example, in a text description of "the server CPU temperature is too high, causing system lag", the large model quickly identifies that this information belongs to different entity types based on the corresponding entity category labels such as "server", "CPU temperature", and "system lag"; according to the relationship category label corresponding to "causes", it determines the relationship between different entities, avoiding incorrect classification and recognition.

[0054] Input the concept object information, data attribute information, object attribute information and their corresponding tags, as well as historical operation and maintenance fault information into a preset large model, so that the large model can perform entity recognition operations and relationship extraction operations on these data. Among them, the entity recognition operation is for the large model to find entities belonging to concept object information and data attribute information from the historical operation and maintenance fault information. For example, in a text "On October 1, 2024, server S1 had a problem of excessive CPU usage", through entity recognition, the large model can find entities such as "October 1, 2024" (corresponding to the entity category of "time") and "excessive CPU usage" (corresponding to the entity category of "fault phenomenon"). The relationship extraction operation is to let the large model extract the relationship represented by the object attribute information from the text. In the above example, the large model can extract the "occurrence" relationship between "server S1" and "excessive CPU usage". Based on the entity recognition operation and the relationship extraction operation, the large model will output the extracted operation and maintenance fault information in the form of triples to obtain the target triples. Among them, a triple usually consists of (entity 1, relationship, entity 2), such as (server S1, occurrence, excessive CPU usage), (October 1, 2024, associated with, the problem of server S1 having excessive CPU usage). These target triples are a structured representation of the historical operation and maintenance fault information, clearly showing the entities related to the operation and maintenance faults and the relationships between them, providing an important data basis for subsequent construction of an operation and maintenance fault handling knowledge graph, fault analysis and diagnosis, etc. By converting complex historical operation and maintenance fault information into a simple triple form, it enables the computer to more conveniently store, query and reason about this information, thereby improving the efficiency and accuracy of operation and maintenance work.

[0055] Among them, a large model using the Transformer architecture can extract the fault diagnosis knowledge defined by the ontology in the field of operation and maintenance fault diagnosis and processing from the historical operation and maintenance fault information through entity recognition and relationship extraction, such as (Jinan Cloud Center, occurrence, high broadband usage), to generate an operation and maintenance fault diagnosis and processing knowledge graph. Among them, the large model based on the Transformer architecture can be a BERT (Bidirectional Encoder Representations from Transformers) discriminative large model and a GPT (Generative Pre-trained Transformer) generative large model.

[0056] Further, an operation and maintenance fault handling knowledge graph is constructed based on the target triples. It can be understood that the knowledge graph consists of nodes and edges. In the operation and maintenance fault handling knowledge graph, the entities in the target triples correspond to the nodes of the knowledge graph, and the relationships correspond to the edges connecting the nodes. Each target triple jointly constructs a complex knowledge network structure.

[0057] Step S12: Perform dependency syntactic analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntactic tree, and perform standardization processing on the dependency syntactic tree based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntactic tree. Use the first dependency syntactic tree to obtain a fault handling question-answer template library; the fault handling question-answer template library includes a target dependency syntactic tree after removing entity information from the first dependency syntactic tree and a corresponding question-answer template.

[0058] In this embodiment, as shown in Figure 2 First, grammatical analysis technology can be used for the historical operation and maintenance fault information to perform Chinese word segmentation and part-of-speech tagging on the data in the historical operation and maintenance fault information, use a preset large model to perform entity recognition on the word segmentation results, and based on the above word segmentation results, part-of-speech tagging results and entity recognition results, use dependency syntactic analysis to generate a dependency syntactic tree. Then, perform standardization processing on the dependency syntactic tree based on the operation and maintenance fault handling knowledge graph in step S11 to obtain a first dependency syntactic tree, which may include: for the dependency syntactic tree, use entity matching technology based on similarity calculation to determine the matching subgraph with the highest similarity to the dependency syntactic tree from the operation and maintenance fault handling knowledge graph; then complete entity alignment between the dependency syntactic tree and the matching subgraph based on rule extraction technology, and use the dependency syntactic tree after entity alignment as the first dependency syntactic tree.

[0059] It can be understood that among the dependency syntactic trees obtained after performing dependency syntactic analysis on the historical operation and maintenance fault information, there may be dependency syntactic trees with similar contents. Therefore, the operation and maintenance fault handling knowledge graph can be used to perform standardization processing on the dependency syntactic tree to obtain a first dependency syntactic tree to unify the dependency syntactic trees with similar contents. At the same time, the expressions of the entities in the dependency syntactic tree can also be aligned with the expressions of the entities in the operation and maintenance fault handling knowledge graph. For example, the dependency syntactic tree may use the specific device name "Server A", while the knowledge graph uses the more general name "server". After entity alignment, the device name can be changed to the general name.

[0060] In this way, the first dependency syntactic tree obtained after completing entity alignment is more standardized and normalized than the original dependency syntactic tree, providing a more reliable basis for subsequent generation of a fault handling question-answer template library, etc., and helping to improve the efficiency and accuracy of fault handling.

[0061] Further, the fault handling Q&A template library can be obtained by using the first dependency syntax tree, including:

[0062] Perform entity information removal operation on the first dependency syntax tree to obtain a target dependency syntax tree, and determine Q&A templates based on the concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntax tree; then determine the fault handling Q&A template library based on the target dependency syntax tree and the corresponding Q&A templates.

[0063] It should be noted that the first dependency grammar tree contains specific entity information. The entity information removal operation is to replace these specific entities with more general placeholders. For example, the sentence corresponding to the first dependency grammar tree is "The hard disk of server A has an over - high temperature, resulting in data loss". After removing the entity information in this sentence, the sentence corresponding to the target dependency grammar tree is "[Device]'s [Component] has [Fault Phenomenon], resulting in [Business Impact]". Then, based on the concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntax tree, determine Q&A templates. For example, the determined Q&A template framework is "When [Device]'s [Component] has [Fault Phenomenon] resulting in [Business Impact]". Then, on the basis of this Q&A template framework, combined with the actual operation and maintenance requirements, transform it into a Q&A form. Since the ultimate goal is to obtain a solution to the fault, a specific position can be set in the template and represented by the target wildcard "?s" for the alarm solution, that is, "When [Device]'s [Component] has [Fault Situation] resulting in [Business Impact],?s". In this way, the Q&A template corresponding to the target dependency syntax tree is obtained, and then the fault handling Q&A template library is obtained based on the target dependency syntax tree and the corresponding Q&A templates. Among them, any template in the fault handling Q&A template library can be applied to multiple similar fault scenarios. As long as the specific entities and situations are filled in the corresponding positions, questions can be raised for different faults and solutions can be obtained.

[0064] Step S13: Determine a second dependency syntax tree based on the operation and maintenance fault information to be processed sent by the user terminal, and use the target dependency syntax tree to determine the Q&A template corresponding to the second dependency syntax tree from the fault handling Q&A template library, and instantiate the Q&A template according to the operation and maintenance fault information to be processed to obtain a semantic query graph.

[0065] In this embodiment, refer to Figure 3As shown, determining a second dependency syntax tree based on the operation and maintenance fault information to be processed sent by the client, and using the target dependency syntax tree to determine the corresponding Q&A template from the fault handling Q&A template library may include: First, obtain the operation and maintenance fault information to be processed sent by the client, determine the second dependency syntax tree corresponding to the operation and maintenance fault information to be processed, and obtain the second target dependency syntax tree after removing entity information from the second dependency syntax tree; Then, determine the target dependency syntax tree corresponding to the second target dependency syntax tree from the fault handling Q&A template library. For example, by comparing the characteristics such as the structure and word relationships of the second target dependency syntax tree with the similarity degree of the target dependency syntax tree in the template library, find the most matching target dependency syntax tree to obtain the corresponding Q&A template through similarity retrieval. Among them, first, Chinese word segmentation and part-of-speech tagging can be performed on the data in the operation and maintenance fault information to be processed, use a preset large model to perform entity recognition on the word segmentation results, and based on the above word segmentation results, part-of-speech tagging results and entity recognition results, use dependency syntax analysis to generate a second dependency syntax tree.

[0066] Further, determine the target template that does not use the target wildcard in the Q&A template, and instantiate the target template based on the operation and maintenance fault information to be processed to obtain the semantic query graph corresponding to the Q&A template.

[0067] It should be noted that determining the target template that does not use the target wildcard in the Q&A template, that is, determining the part in the Q&A template that needs to be filled according to the operation and maintenance fault information to be processed, and then instantiating the target template based on the operation and maintenance fault information. Taking "The hard disk of server A has read and write errors, resulting in abnormal data transmission" as an example, fill "server A" into "[device]", "hard disk" into "[component]", "read and write errors" into "[fault phenomenon]", and "abnormal data transmission" into "[business impact]" to obtain the instantiated question "What measures should be taken when the hard disk of server A has read and write errors resulting in abnormal data transmission?". Finally, convert this instantiated question into a semantic query graph. The semantic query graph graphically shows the entities and relationships in the question, facilitating computer understanding and further processing, so as to be able to more accurately query relevant fault solutions from the knowledge base or other data sources.

[0068] Step S14: Determine the target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and obtain the operation and maintenance fault handling result based on the target subgraph.

[0069] In this embodiment, first, the target subgraph corresponding to the semantic query graph can be determined from the operation and maintenance fault handling knowledge graph, and the similarity between the semantic query graph and each target subgraph can be determined. Then, the target subgraph with the highest similarity is selected as the final target subgraph, and the alarm solution corresponding to the final target subgraph is used as the operation and maintenance fault handling result corresponding to the operation and maintenance fault information to be processed.

[0070] Among them, the subgraph matching technology based on graph neural network can be used to match and obtain multiple target subgraphs from the operation and maintenance fault handling knowledge graph, and then calculate the similarity between the semantic query subgraph and the target subgraph, and output the content of "alarm solution" in the target subgraph with the highest similarity as the operation and maintenance fault handling result. Among them, the similarity calculation formula between the semantic query subgraph and the target subgraph is as follows:

[0071] ;

[0072] Among them, and are the semantic query subgraph and the target subgraph respectively, represents the set of entity nodes in the semantic query subgraph, represents the th entity node, is a custom coefficient, is a function for calculating the entity similarity between the semantic query subgraph and the target subgraph.

[0073] As can be seen from the above, in this embodiment, by constructing an ontology in the field of operation and maintenance fault handling, and based on the ontology in the field of operation and maintenance fault handling and a preset large model, a complete and reliable operation and maintenance fault handling knowledge graph is obtained for the effective management and efficient reuse of historical operation and maintenance fault information; by using the constructed operation and maintenance fault handling knowledge graph, combined with grammar analysis technology and rule extraction technology, a fault handling Q&A template library is obtained, so that based on the fault handling Q&A template library, the operation and maintenance fault information to be processed can be quickly mapped to the corresponding Q&A template to generate a semantic query graph, and through subgraph matching technology, the target subgraph corresponding to the semantic query graph is determined from the operation and maintenance fault handling knowledge graph, and the operation and maintenance fault handling result is obtained based on the target subgraph, so as to realize the instant response and efficient processing of operation and maintenance faults, and can identify the root cause of faults more quickly and accurately. At the same time, in this embodiment, through the automated and intelligent operation and maintenance fault handling method, not only the dependence on manual experience in operation and maintenance fault diagnosis work is reduced, the operation and maintenance work efficiency is improved, but also the operation and maintenance fault decision-making process can be optimized in a data-driven manner, which helps to build a more stable and reliable operation and maintenance fault management information system, laying a solid technical foundation for the digital transformation of enterprise operation and maintenance.

[0074] See Figure 4As shown below, taking the overall schematic diagram of the operation and maintenance fault handling method based on the large model and knowledge graph as an example, the technical solutions in this application will be described.

[0075] This embodiment is divided into a historical operation and maintenance fault case management service and an operation and maintenance fault instant response handling service.

[0076] In the historical operation and maintenance fault case management service, first, obtain historical operation and maintenance fault information, construct an ontology for the operation and maintenance fault handling field based on the historical operation and maintenance fault information to complete knowledge modeling; then input the historical operation and maintenance fault information and the ontology for the operation and maintenance fault handling field into a preset large model to obtain the target triples output by the preset large model, and construct an operation and maintenance fault handling knowledge graph based on the target triples; further, perform Chinese word segmentation and part-of-speech tagging on the data in the historical operation and maintenance fault information, use the preset large model to perform entity recognition on the word segmentation results, and based on the above word segmentation results, part-of-speech tagging results, and entity recognition results, use dependency syntax analysis to generate a dependency syntax tree, and then perform standardization processing on the dependency syntax tree based on the operation and maintenance fault handling knowledge graph to obtain the first dependency syntax tree; perform an entity information removal operation on the first dependency syntax tree to obtain the target dependency syntax tree, and determine a question and answer template based on the concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntax tree, and finally determine a fault handling question and answer template library based on the target dependency syntax tree and the corresponding question and answer template to complete knowledge extraction.

[0077] In the operation and maintenance fault instant response handling service, first, determine a second dependency syntax tree based on the operation and maintenance fault information to be processed sent by the user terminal, and use the target dependency syntax tree to determine the question and answer template corresponding to the second dependency syntax tree from the fault handling question and answer template library, then instantiate the question and answer template according to the operation and maintenance fault information to be processed to obtain a semantic query graph, determine the target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and finally obtain the operation and maintenance fault handling result based on the target subgraph to realize the automatic response of the operation and maintenance fault handling and complete knowledge reasoning.

[0078] As can be seen from the above, this embodiment realizes the effective integration and efficient reuse of historical operation and maintenance fault cases based on the historical operation and maintenance fault case management service, further improving the utilization rate of resources; at the same time, according to the operation and maintenance fault instant response handling service, it realizes the intelligence and automation of the operation and maintenance fault handling process, can more quickly and accurately identify the root cause of the fault, recommend reasonable solutions, so as to reduce the manual operation and maintenance management cost and improve the operation and maintenance work efficiency.

[0079] See Figure 5 As shown, this application embodiment also discloses an operation and maintenance fault handling device based on a large model and a knowledge graph, including:

[0080] A knowledge graph determination module 11, configured to construct an ontology for operation and maintenance fault handling in the field based on historical operation and maintenance fault information, and input the historical operation and maintenance fault information and the ontology for operation and maintenance fault handling in the field into a preset large model to obtain target triples output by the preset large model, and obtain an operation and maintenance fault handling knowledge graph based on the target triples; the ontology for operation and maintenance fault handling in the field is used to describe the types, properties of each data in the historical operation and maintenance fault information, and the relationships between the data;

[0081] A template library determination module 12, configured to perform dependency syntactic analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntactic tree, perform standardization processing on the dependency syntactic tree based on the operation and maintenance fault handling knowledge graph to obtain a first dependency syntactic tree, and use the first dependency syntactic tree to obtain a fault handling question and answer template library; the fault handling question and answer template library includes a target dependency syntactic tree after removing entity information from the first dependency syntactic tree and a corresponding question and answer template;

[0082] A semantic query graph determination module 13, configured to determine a second dependency syntactic tree based on the operation and maintenance fault information to be processed sent by the user side, use the target dependency syntactic tree to determine a question and answer template corresponding to the second dependency syntactic tree from the fault handling question and answer template library, and instantiate the question and answer template according to the operation and maintenance fault information to be processed to obtain a semantic query graph;

[0083] A result determination module 14, configured to determine a target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, and obtain an operation and maintenance fault handling result based on the target subgraph.

[0084] As can be seen from the above, in this application, an ontology for operation and maintenance fault handling in the field is constructed, and an operation and maintenance fault handling knowledge graph is obtained based on the ontology for operation and maintenance fault handling in the field and a preset large model, which is used for effective management and efficient reuse of historical operation and maintenance fault information; the operation and maintenance fault handling knowledge graph is used to perform standardization processing on the dependency syntactic tree to obtain a first dependency syntactic tree, and the first dependency syntactic tree is used to obtain a fault handling question and answer template library, so that based on the fault handling question and answer template library, for the operation and maintenance fault information to be processed, it can be quickly mapped to the corresponding question and answer template to generate a semantic query graph, and through subgraph matching technology, a target subgraph corresponding to the semantic query graph is determined from the operation and maintenance fault handling knowledge graph, and an operation and maintenance fault handling result is obtained based on the target subgraph, so as to achieve instant response and efficient processing of operation and maintenance faults, be able to identify the root cause of faults more quickly and accurately, and at the same time reduce the dependence on manual experience in operation and maintenance fault diagnosis work and improve the operation and maintenance work efficiency.

[0085] In some specific embodiments, the knowledge graph determination module 11 includes:

[0086] An ontology construction unit, configured to extract concept object information, data attribute information, and object attribute information from historical operation and maintenance fault information, and construct an ontology in the field of operation and maintenance fault handling based on the concept object information, the data attribute information, and the object attribute information;

[0087] Among them, the concept object information includes alarm events, alarm names, and alarm solutions; the data attribute information includes alarm times and IP addresses; the object attribute information includes first attribute information representing the occurrence of alarms, second attribute information representing the association between different alarms, and third attribute information representing alarm handling methods.

[0088] In some specific embodiments, the knowledge graph determination module 11 includes:

[0089] A data input unit, configured to assign entity category labels to the concept object information and the data attribute information, assign relationship category labels to the object attribute information, and input the concept object information, the data attribute information, the object attribute information, the corresponding labels, and the historical operation and maintenance fault information into a preset large model;

[0090] A triple acquisition unit, configured to perform entity recognition operations and relationship extraction operations based on the preset large model, so as to extract operation and maintenance fault information defined by the operation and maintenance fault handling ontology from the historical operation and maintenance fault information, and output it in the form of triples to obtain target triples.

[0091] In some specific embodiments, the template library determination module 12 includes:

[0092] A subgraph determination unit, configured to, for the dependency syntactic tree, use entity matching technology based on similarity calculation to determine a matching subgraph with the highest similarity to the dependency syntactic tree from the operation and maintenance fault handling knowledge graph;

[0093] A first syntactic tree determination unit, configured to complete entity alignment between the dependency syntactic tree and the matching subgraph based on rule extraction technology, and use the dependency syntactic tree after entity alignment as the first dependency syntactic tree;

[0094] A first template determination unit, configured to perform an entity information removal operation on the first dependency syntactic tree to obtain a target dependency syntactic tree, and determine a question and answer template based on concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntactic tree;

[0095] A template library determination unit, configured to determine a fault handling question and answer template library based on the target dependency syntactic tree and the corresponding question and answer template.

[0096] In some specific embodiments, the semantic query graph determination module 13 includes:

[0097] A second syntactic tree determination unit, configured to obtain the to-be-processed operation and maintenance fault information sent by the user terminal, determine the second dependency syntactic tree corresponding to the to-be-processed operation and maintenance fault information, and obtain a second target dependency syntactic tree after removing entity information from the second dependency syntactic tree;

[0098] A second template determination unit, configured to determine a target dependency syntactic tree corresponding to the second target dependency syntactic tree from the fault handling Q&A template library, so as to obtain a corresponding Q&A template.

[0099] In some specific embodiments, the semantic query graph determination module 13 includes:

[0100] A semantic query graph determination unit, configured to determine a target template that does not use a target wildcard representation in the Q&A template, and instantiate the target template based on the to-be-processed operation and maintenance fault information to obtain a semantic query graph corresponding to the Q&A template.

[0101] In some specific embodiments, the result determination module 14 includes:

[0102] A target subgraph determination unit, configured to determine a target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, determine the similarity between the semantic query graph and each target subgraph, and select the target subgraph with the highest similarity as the final target subgraph;

[0103] A result acquisition unit, configured to use the alarm solution corresponding to the final target subgraph as the operation and maintenance fault handling result corresponding to the to-be-processed operation and maintenance fault information.

[0104] Furthermore, an electronic device is also disclosed in an embodiment of the present application. Figure 6 It is a structural diagram of an electronic device 20 shown according to an exemplary embodiment, and the content in the figure should not be considered as any limitation to the scope of use of the present application.

[0105] Figure 6 It is a structural schematic diagram of an electronic device 20 provided by an embodiment of the present application. The electronic device 20 may specifically include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. Among them, the memory 22 is used to store a computer program, and the computer program is loaded and executed by the processor 21 to implement the relevant steps in the operation and maintenance fault handling method based on a large model and a knowledge graph disclosed in any of the foregoing embodiments. In addition, the electronic device 20 in this embodiment may specifically be an electronic computer.

[0106] In this embodiment, the power supply 23 is used to provide operating voltages for the various hardware devices on the electronic device 20; the communication interface 24 can create a data transmission channel between the electronic device 20 and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and specific limitations thereof are not provided herein; the input / output interface 25 is used to obtain external input data or output data to the outside, and the specific interface type thereof can be selected according to specific application requirements, and specific limitations are not provided herein.

[0107] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, a random access memory, a magnetic disk, an optical disc, etc., and the resources stored thereon can include an operating system 221, a computer program 222, etc., and the storage method can be transient storage or permanent storage.

[0108] Among them, the operating system 221 is used to manage and control the various hardware devices and the computer program 222 on the electronic device 20, and it can be Windows Server, Netware, Unix, Linux, etc. The computer program 222, in addition to including a computer program capable of implementing the operation and maintenance fault handling method based on the large model and knowledge graph executed by the electronic device 20 disclosed in any of the foregoing embodiments, may further include a computer program capable of performing other specific tasks.

[0109] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the operation and maintenance fault handling method based on the large model and knowledge graph disclosed above. For the specific steps of this method, reference can be made to the corresponding content disclosed in the foregoing embodiments, and details are not described herein again.

[0110] In this specification, the various embodiments are described in a progressive manner, and the key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0111] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0112] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented directly in hardware, in software modules executed by a processor, or in a combination thereof. The software modules may be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well known in the art.

[0113] Finally, it should also be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0114] The technical solutions provided in this application have been introduced in detail above. Specific examples are used herein to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A maintenance fault handling method based on a large model and knowledge graph, characterized in that: include: Construct an operation and maintenance fault processing domain ontology based on historical operation and maintenance fault information, and input the historical operation and maintenance fault information and the operation and maintenance fault processing domain ontology into a preset large model to obtain a target triplet output by the preset large model, and obtain an operation and maintenance fault processing knowledge graph based on the target triplet; The operation and maintenance fault processing domain ontology is used to describe the type and nature of each data in the historical operation and maintenance fault information and the relationship between the data; Performing dependency syntax analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntax tree, and performing standardization processing on the dependency syntax tree based on the operation and maintenance fault processing knowledge graph to obtain a first dependency syntax tree, and using the first dependency syntax tree to obtain a fault processing question and answer template library; The fault handling question and answer template library includes a target dependency syntax tree after removing entity information from the first dependency syntax tree and a corresponding question and answer template; Determine a second dependency syntax tree based on the pending operation and maintenance fault information sent by the user terminal, and use the target dependency syntax tree to determine the question and answer template corresponding to the second dependency syntax tree from the fault processing question and answer template library, and instantiate the question and answer template according to the pending operation and maintenance fault information to obtain a semantic query graph; A target subgraph corresponding to the semantic query graph is determined from the operation and maintenance fault handling knowledge graph, and an operation and maintenance fault handling result is obtained based on the target subgraph.

2. The operation and maintenance fault handling method based on a large model and a knowledge graph according to claim 1 is characterized in that: The construction of the operation and maintenance fault processing domain ontology based on historical operation and maintenance fault information includes: Extracting conceptual object information, data attribute information and object attribute information from historical operation and maintenance fault information, and constructing an operation and maintenance fault processing domain ontology based on the conceptual object information, the data attribute information and the object attribute information; Among them, the conceptual object information includes alarm events, alarm names, and alarm solutions; the data attribute information includes alarm time and IP address; the object attribute information includes first attribute information representing the occurrence of an alarm, second attribute information representing the relationship between different alarms, and third attribute information representing the alarm processing method.

3. The operation and maintenance fault handling method based on a large model and a knowledge graph according to claim 2 is characterized in that: The inputting of the historical operation and maintenance fault information and the operation and maintenance fault processing domain ontology into the preset large model to obtain the target triplet output by the preset large model includes: Assign entity category labels to the concept object information and the data attribute information, assign relationship category labels to the object attribute information, and input the concept object information, the data attribute information, the object attribute information and the corresponding labels, and the historical operation and maintenance fault information into a preset large model; Based on the preset large model, entity recognition operations and relationship extraction operations are performed to extract the operation and maintenance fault information defined by the operation and maintenance fault processing domain ontology from the historical operation and maintenance fault information, and output it in the form of triples to obtain target triples.

4. The operation and maintenance fault handling method based on a large model and a knowledge graph according to any one of claims 2 to 3, characterized in that: The step of normalizing the dependency syntax tree based on the operation and maintenance fault processing knowledge graph to obtain a first dependency syntax tree includes: For the dependency syntax tree, determine a matching subgraph with the highest similarity to the dependency syntax tree from the operation and maintenance fault handling knowledge graph using an entity matching technology based on similarity calculation; Based on the rule extraction technology, entity alignment between the dependency syntax tree and the matching subgraph is completed, and the dependency syntax tree after entity alignment is used as the first dependency syntax tree; Correspondingly, the method of obtaining a fault handling question and answer template library by using the first dependency syntax tree includes: Performing an entity information removal operation on the first dependency syntax tree to obtain a target dependency syntax tree, and determining a question-answering template based on concept object class nodes, data attribute class nodes, and object attribute edge relationships in the target dependency syntax tree; A fault handling question and answer template library is determined based on the target dependency syntax tree and the corresponding question and answer template.

5. The operation and maintenance fault handling method based on a large model and a knowledge graph according to claim 1 is characterized in that: The determining of the second dependency syntax tree based on the pending operation and maintenance fault information sent by the user terminal, and determining the question and answer template corresponding to the second dependency syntax tree from the fault handling question and answer template library using the target dependency syntax tree, includes: Obtaining the pending operation and maintenance fault information sent by the user terminal, determining a second dependency syntax tree corresponding to the pending operation and maintenance fault information, and removing entity information from the second dependency syntax tree to obtain a second target dependency syntax tree; A target dependency syntax tree corresponding to the second target dependency syntax tree is determined from the fault handling question and answer template library to obtain a corresponding question and answer template.

6. The operation and maintenance fault handling method based on a large model and knowledge graph according to claim 1 is characterized in that: The instantiating the question-answer template according to the operation and maintenance fault information to be processed to obtain a semantic query graph includes: A target template that is not represented by a target wildcard in the question and answer template is determined, and the target template is instantiated based on the operation and maintenance fault information to be processed to obtain a semantic query graph corresponding to the question and answer template.

7. The operation and maintenance fault handling method based on a large model and knowledge graph according to claim 1 is characterized in that: The step of determining a target subgraph corresponding to the semantic query graph from the operation and maintenance fault processing knowledge graph, and obtaining an operation and maintenance fault processing result based on the target subgraph, includes: Determine the target subgraph corresponding to the semantic query graph from the operation and maintenance fault handling knowledge graph, determine the similarity between the semantic query graph and each of the target subgraphs, and select the target subgraph with the highest similarity as the final target subgraph; The corresponding alarm solution in the final target subgraph is used as the operation and maintenance fault processing result corresponding to the operation and maintenance fault information to be processed.

8. An operation and maintenance fault handling device based on a large model and knowledge graph, characterized in that: include: A knowledge graph determination module is used to construct an operation and maintenance fault processing domain ontology based on historical operation and maintenance fault information, and input the historical operation and maintenance fault information and the operation and maintenance fault processing domain ontology into a preset large model to obtain a target triple output by the preset large model, and obtain an operation and maintenance fault processing knowledge graph based on the target triple; The operation and maintenance fault processing domain ontology is used to describe the type and nature of each data in the historical operation and maintenance fault information and the relationship between the data; A template library determination module, used to perform dependency syntax analysis on the historical operation and maintenance fault information to obtain a corresponding dependency syntax tree, and to perform standardization processing on the dependency syntax tree based on the operation and maintenance fault processing knowledge graph to obtain a first dependency syntax tree, and to obtain a fault processing question and answer template library using the first dependency syntax tree; The fault handling question and answer template library includes a target dependency syntax tree after removing entity information from the first dependency syntax tree and a corresponding question and answer template; A semantic query graph determination module, configured to determine a second dependency syntax tree based on the pending operation and maintenance fault information sent by the user terminal, and to determine a question and answer template corresponding to the second dependency syntax tree from the fault processing question and answer template library using the target dependency syntax tree, and to instantiate the question and answer template according to the pending operation and maintenance fault information to obtain a semantic query graph; The result determination module is used to determine the target subgraph corresponding to the semantic query graph from the operation and maintenance fault processing knowledge graph, and obtain the operation and maintenance fault processing result based on the target subgraph.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, used to execute the computer program to implement the operation and maintenance fault handling method based on a large model and a knowledge graph as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store a computer program, which, when executed by a processor, implements the operation and maintenance fault handling method based on a large model and a knowledge graph as described in any one of claims 1 to 7.

Images