Power grid fault disposal knowledge base management system

Abstract

The invention provides a power grid fault processing knowledge base management system, and relates to the technical field of power grid fault processing, and the system comprises a data storage module which is used for obtaining and storing related knowledge data of the power grid industry; the knowledge graph construction module is used for constructing a power grid knowledge graph based on the related knowledge data of the power grid industry; the query retrieval module comprises a statement generation unit and a scheme retrieval unit; the statement generation unit is used for generating a graph query statement based on a power grid fault problem proposed by a user; and the scheme retrieval unit is used for retrieving a first power grid fault handling scheme from the knowledge graph based on the graph query statement. According to the system, the power grid knowledge graph is constructed, and the query retrieval module is trained through the content in the power grid knowledge graph, so that the system can quickly and accurately retrieve the corresponding power grid fault handling scheme through the power grid fault problem proposed by the user.

Description

Technical Field

[0001] This invention relates to the field of power grid fault handling technology, and in particular to a power grid fault handling knowledge base management system. Background Technology

[0002] With the continuous development of society and the expansion of the power grid, the types of power grid faults are becoming increasingly complex. This places higher demands on the professional capabilities of maintenance personnel when repairing faulty power grids. Currently, power grid fault handling mainly relies on maintenance personnel's understanding of existing fault handling solutions. Maintenance personnel select the most suitable solution from existing solutions based on the type of power grid fault and then repair the faulty grid according to the solution. However, due to potential biases in their understanding of power grid fault handling solutions, maintenance personnel may be unable to determine the most suitable solution based on the specific fault. Furthermore, because there are many types of power grid faults and corresponding solutions, the efficiency of searching for solutions is low. Therefore, there are technical problems of low efficiency and low accuracy in searching for power grid fault handling solutions. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide a power grid fault handling knowledge base management system that can solve the technical problems of low retrieval efficiency and low accuracy in the prior art when searching for power grid fault handling solutions.

[0004] To achieve the above objectives, the technical solutions adopted in the embodiments of the present invention are as follows:

[0005] In a first aspect, embodiments of the present invention provide a power grid fault handling knowledge base management system, comprising:

[0006] Data storage module: used to acquire and store knowledge data related to the power grid industry; wherein, the knowledge data related to the power grid industry includes power grid fault problems and corresponding power grid fault handling solutions;

[0007] Knowledge graph construction module: used to construct a power grid knowledge graph based on the power grid industry-related knowledge data;

[0008] The query and retrieval module includes a statement generation unit and a solution retrieval unit. The statement generation unit generates a graph query statement based on the power grid fault problem proposed by the user. The solution retrieval unit retrieves a first power grid fault handling solution from the power grid knowledge graph based on the graph query statement. The query and retrieval module is trained using a training dataset. The training dataset includes power grid fault problems labeled with graph query statements and a power grid knowledge graph.

[0009] Furthermore, this embodiment of the invention provides a first possible implementation of the first aspect, wherein constructing a power grid knowledge graph based on the power grid industry-related knowledge data includes:

[0010] Determine the entities and relationships between them in the power grid fault problem and its corresponding power grid fault handling plan;

[0011] Based on the entities and the relationships between them, triples are constructed, and a power grid knowledge graph is built based on the triples.

[0012] Furthermore, this embodiment of the invention provides a second possible implementation of the first aspect, which further includes: a database construction module and a database filtering module;

[0013] The database construction module is used to input the power grid industry-related knowledge data into the natural language processing module for processing, and generate a power grid vector database.

[0014] The database filtering module is used to filter multiple fault handling solution vectors from the power grid vector database based on the power grid fault problem proposed by the user.

[0015] Furthermore, this embodiment of the invention provides a third possible implementation of the first aspect, wherein the step of inputting the power grid industry-related knowledge data into a natural language processing module for processing to generate a power grid vector database includes:

[0016] Based on the natural language processing module, the power grid fault problem and its corresponding power grid fault handling solution are processed by text segmentation to obtain multiple problem text blocks and their corresponding fault handling solution text blocks.

[0017] Each of the aforementioned problem text blocks and its corresponding fault handling solution text blocks are converted into a first problem vector and its corresponding fault handling solution vector, respectively, to obtain a power grid vector database.

[0018] Furthermore, this embodiment of the invention provides a fourth possible implementation of the first aspect, wherein the step of selecting multiple fault handling scheme vectors from the power grid vector database based on the power grid fault problem proposed by the user includes:

[0019] Determine the corresponding second problem vector based on the power grid fault problem proposed by the user;

[0020] Calculate the cosine similarity between the second problem vector and each of the first problem vectors, determine the first problem vector and its corresponding fault handling solution vector when the cosine similarity is greater than a preset similarity threshold, and select multiple fault handling solution vectors from the power grid vector database.

[0021] Furthermore, this embodiment of the invention provides a fifth possible implementation of the first aspect, wherein retrieving the first power grid fault handling solution from the power grid knowledge graph based on the graph query statement includes:

[0022] Determine each second power grid fault handling scheme corresponding to multiple fault handling scheme vectors selected from the power grid vector database;

[0023] The content corresponding to each of the second power grid fault handling schemes in the power grid knowledge graph is determined, and the first power grid fault handling scheme is retrieved from it based on the graph query statement.

[0024] Furthermore, this embodiment of the invention provides a sixth possible implementation of the first aspect, which further includes: a training module;

[0025] A sample dataset is constructed based on the fault problems of each power grid and their corresponding graph query statements, as well as the fault problems of each power grid and their corresponding power grid fault handling solutions in the power grid knowledge graph.

[0026] The sample dataset is input into a large language model for training to obtain the query retrieval module.

[0027] Furthermore, this embodiment of the invention provides a seventh possible implementation of the first aspect, which further includes: a performance testing module;

[0028] The sample dataset also includes a test dataset;

[0029] The performance testing module is used to input the test dataset into the query retrieval module to obtain performance testing parameters to evaluate the performance of the query retrieval module; wherein, the performance testing parameters include: precision, recall and F1 score.

[0030] Furthermore, this embodiment of the invention provides an eighth possible implementation of the first aspect, which further includes: a human-computer interaction module;

[0031] The human-computer interaction module is used to communicate with the user and to visually display the first power grid fault handling solution retrieved based on the power grid fault problem raised by the user.

[0032] Furthermore, this embodiment of the invention provides a ninth possible implementation of the first aspect, which further includes: an update module;

[0033] The update module is used to receive power grid industry-related knowledge data in real time, and update the power grid knowledge graph and power grid vector database based on the received power grid industry-related knowledge data in real time.

[0034] This invention provides a power grid fault handling knowledge base management system, which includes: a data storage module for acquiring and storing power grid industry-related knowledge data, wherein the power grid industry-related knowledge data includes power grid fault problems and their corresponding power grid fault handling solutions; a knowledge graph construction module for constructing a power grid knowledge graph based on the power grid industry-related knowledge data; and a query and retrieval module including a statement generation unit and a solution retrieval unit, wherein the statement generation unit generates a graph query statement based on a power grid fault problem proposed by a user; the solution retrieval unit retrieves a first power grid fault handling solution from the power grid knowledge graph based on the graph query statement; the query and retrieval module is trained using a training dataset; the training dataset includes: power grid fault problems labeled with graph query statements and a power grid knowledge graph. The present invention constructs a power grid knowledge graph by identifying power grid fault problems and their corresponding power grid fault handling solutions. This enables the unification of multi-source, heterogeneous power grid industry-related knowledge data. The query and retrieval module has been trained using the power grid knowledge graph and power grid fault problems marked with graph query statements. The power grid knowledge graph enhances the context awareness and logical reasoning capabilities of the query and retrieval module. By receiving the power grid fault problem proposed by the user, the query and retrieval module can quickly and accurately obtain the graph query statement and the first power grid fault handling solution retrieved from the power grid knowledge graph using the graph query statement, thereby improving the efficiency and accuracy of retrieving power grid fault handling solutions.

[0035] Other features and advantages of the embodiments of the present invention will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above in the embodiments of the present invention.

[0036] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 This diagram illustrates the main modules of a power grid fault handling knowledge base management system provided in an embodiment of the present invention.

[0039] Figure 2This diagram illustrates a unit schematic of a query and retrieval module in a power grid fault handling knowledge base management system provided in an embodiment of the present invention.

[0040] Figure 3 This diagram illustrates the processing steps of a knowledge graph construction module in a power grid fault handling knowledge base management system provided by an embodiment of the present invention.

[0041] Figure 4 This diagram illustrates all the modules in a power grid fault handling knowledge base management system provided by an embodiment of the present invention.

[0042] Figure 5 This diagram illustrates the processing steps of the database construction module in a power grid fault handling knowledge base management system provided by an embodiment of the present invention.

[0043] Figure 6 This diagram illustrates the processing steps of a database filtering module in a power grid fault handling knowledge base management system provided by an embodiment of the present invention.

[0044] Figure 7 The diagram illustrates the processing steps of a solution retrieval unit in a power grid fault handling knowledge base management system provided by an embodiment of the present invention. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be described below in conjunction with the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0046] This embodiment provides a power grid fault handling knowledge base management system. For details, please refer to... Figure 1 The diagram shows the main modules of a power grid fault handling knowledge base management system. The system includes: a data storage module 11, a knowledge graph construction module 12, and a query and retrieval module 13.

[0047] The data storage module 11 is used to acquire and store knowledge data related to the power grid industry; among which, the knowledge data related to the power grid industry includes power grid fault problems and their corresponding power grid fault handling solutions;

[0048] Data storage module 11 acquires relevant knowledge data about the power grid industry, preprocesses the data, and ultimately determines power grid fault problems and corresponding fault handling solutions. Specifically, data storage module 11 collects a large amount of relevant knowledge data about the power grid industry, including: power dispatch logs, covering the operating status of voltage levels and equipment types in various power grids; maintenance procedure documents for the national power system, which contain standard operating procedures and safety precautions; power grid fault analysis reports and power grid fault handling solutions. During the acquisition of relevant knowledge data about the power grid industry, various types of files exist (such as power grid knowledge documents in PDF format, power grid documents in TXT / JSON / CSV formats). Therefore, different types of power grid files need to be processed, such as obtaining power grid knowledge data from PDF documents using PDF readers (e.g., PyPDFLoader) or optical character recognition technology, directly obtaining power grid knowledge data from TXT format power grid question-and-answer datasets, and obtaining power grid knowledge data from JSON / CSV format power grid question-and-answer datasets using Python. After obtaining the power grid industry-related knowledge data, it is necessary to perform data cleaning, word segmentation, and standardization to convert unstructured power grid industry-related knowledge data into a standard format, so as to accurately obtain power grid fault problems and their corresponding power grid fault handling solutions.

[0049] Knowledge graph construction module 12 is used to construct a power grid knowledge graph based on power grid industry-related knowledge data;

[0050] By identifying the entities and relationships within power grid industry-related knowledge data (e.g., equipment names, fault types, geographical locations, fault causes, and handling methods), and using entities as nodes and relationships between entities as edges, a power grid knowledge graph is constructed. This knowledge graph can transform multi-source, heterogeneous power grid industry-related knowledge data into a unified semantic network, breaking down data silos. Furthermore, when the power grid knowledge graph is used to train the query and retrieval module 13, it can enhance the model's context awareness and dynamic reasoning capabilities, thereby improving the retrieval accuracy of the query and retrieval module 13.

[0051] See Figure 2 The diagram shows a unit schematic of a query and retrieval module in a power grid fault handling knowledge base management system. The query and retrieval module 13 includes a statement generation unit 131 and a solution retrieval unit 132. The statement generation unit 131 is used to generate a graph query statement based on the power grid fault problem proposed by the user. The solution retrieval unit 132 is used to retrieve the first power grid fault handling solution from the power grid knowledge graph based on the graph query statement. The query and retrieval module 13 is trained using a training dataset. The training dataset includes: power grid fault problems labeled with graph query statements and a power grid knowledge graph.

[0052] The aforementioned query retrieval module 13 mainly includes a large language model (such as ChatGPT, LLAMA, etc.). The semantic understanding capability of the large language model is used to ensure the accuracy of the subsequent retrieval of the first power grid fault handling solution. The query retrieval module 13 is obtained by training the large language model through the training dataset. The statement generation unit 131 generates a graph query statement (that is, a Cypher statement that can be used to retrieve the power grid knowledge graph) based on the power grid fault problem proposed by the user. The solution retrieval unit 132 retrieves the power grid fault handling solution that can solve the power grid fault problem proposed by the user from the power grid knowledge graph by using the Cypher statement.

[0053] The power grid fault handling knowledge base management system provided in this embodiment of the invention constructs a power grid knowledge graph by building power grid fault problems and their corresponding power grid fault handling solutions. This enables the unification of multi-source heterogeneous power grid industry-related knowledge data. The query and retrieval module has been trained using the power grid knowledge graph and power grid fault problems marked with graph query statements. The power grid knowledge graph enhances the context awareness and logical reasoning capabilities of the query and retrieval module. By receiving the power grid fault problems proposed by the user, the query and retrieval module can quickly and accurately obtain the graph query statement and the first power grid fault handling solution retrieved from the power grid knowledge graph by the graph query statement, thereby improving the efficiency and accuracy of retrieving power grid fault handling solutions.

[0054] In one embodiment, see Figure 3 The diagram illustrates the processing steps of a knowledge graph construction module in a power grid fault handling knowledge base management system. This embodiment provides a specific implementation method for constructing a power grid knowledge graph based on relevant knowledge data from the power grid industry.

[0055] Step S301: Determine the entities and relationships between them in the power grid fault problem and its corresponding power grid fault handling plan;

[0056] Step S303: Construct triples based on each entity and the relationships between them, and construct a power grid knowledge graph based on the triples;

[0057] The knowledge graph construction module 12 uses a BERT-based sequence labeling model to extract key entities (including equipment names, fault types, fault causes, and handling methods) from relevant knowledge data in the power grid industry. It identifies relationships between entities through a relation extraction model. Specifically, the constructed power grid knowledge graph uses an attribute graph model, containing two basic elements: nodes and edges. Nodes represent entities (including head and tail entities), and edges represent relationships between entities. The main entity types in the power grid knowledge graph include equipment (such as transformers, circuit breakers, and lines), fault types (such as short circuits, overloads, and insulation breakdown), and handling methods (such as power outage maintenance, load transfer, and emergency power generation). Relationship types between entities include "cause," "applies to," and "located in." The power grid knowledge graph is stored using the Neo4j graph database, supporting the Cypher query language for easy querying and reasoning of complex relationships. This module also implements visualization functionality for the power grid knowledge graph, using the D3.js library to intuitively display the graph structure to users, supporting interactive exploration and node expansion.

[0058] In one embodiment, see Figure 4 The diagram shown is a complete module of a power grid fault handling knowledge base management system. The power grid fault handling knowledge base management system provided in this embodiment also includes: a database construction module 14 and a database filtering module 15.

[0059] Database construction module 14 is used to input power grid industry-related knowledge data into the natural language processing module for processing, and generate a power grid vector database;

[0060] By constructing a unified vector space, the power grid vector database can achieve cross-modal retrieval. It can also improve the semantic understanding capability of the power grid vector database by dynamically adjusting the spatial weights of each vector in the database. Furthermore, when updating the power grid vector database, incremental vector updates can be performed without rebuilding the entire index. The vector space of the power grid vector database can be continuously optimized through an online learning mechanism.

[0061] Database filtering module 15 is used to filter multiple fault handling solution vectors from the power grid vector database based on the power grid fault problems proposed by the user;

[0062] The database filtering module 15 first filters out a portion of fault handling scheme vectors that match the power grid fault problem from the power grid vector database, and then filters out the first power grid fault handling scheme from the power grid fault handling scheme vectors to improve the retrieval accuracy.

[0063] In one embodiment, see Figure 5The diagram illustrates the processing steps of a database construction module in a power grid fault handling knowledge base management system. This embodiment provides a specific implementation method for inputting power grid industry-related knowledge data into a natural language processing module for processing to generate a power grid vector database.

[0064] Step S501: Based on the natural language processing module, perform text segmentation processing on the power grid fault problem and its corresponding power grid fault handling solution to obtain multiple problem text blocks and their corresponding fault handling solution text blocks.

[0065] When constructing the power grid vector database, long documents need to be segmented first. The text splitter in the natural language processing module is used to segment power grid industry-related knowledge data (i.e., power grid fault problems and their corresponding power grid fault handling solutions). Specifically, the text splitting method and the size of the segmented blocks are defined in the text splitter, and the text splitter is used to segment the read content. The specific parameters of the text splitter include: delimiter, text block length, overlap length between text blocks, and length calculation method. After segmenting the power grid fault problems and their corresponding power grid fault handling solutions, multiple problem text blocks and their corresponding fault handling solution text blocks are obtained.

[0066] Step S503: Convert each problem text block and its corresponding fault handling solution text block into each first problem vector and its corresponding fault handling solution vector to obtain the power grid vector database.

[0067] When constructing a power grid vector database, the text needs to be vectorized first. The purpose is to convert multiple problem text blocks and their corresponding fault handling solution text blocks into vector form, and then store the vectorized text in the vector database. Specifically, since large language models can only understand numbers and not text information, the text information must be encoded. The encoding method referred to here is embedding. By using text embedding, multiple problem text blocks and their corresponding fault handling solution text blocks are converted into first problem vectors and their corresponding fault handling solution vectors, respectively. The resulting vectors are stored in a vector database (Milvus vector database can be used, which has efficient similarity search and uses IVF_FLAT index type), and finally the power grid vector database is obtained.

[0068] In one embodiment, see Figure 6 The diagram illustrates the processing steps of a database filtering module in a power grid fault handling knowledge base management system. This embodiment provides a specific implementation method for filtering multiple fault handling solution vectors from a power grid vector database based on power grid fault problems proposed by users.

[0069] Step S601: Determine the corresponding second problem vector based on the power grid fault problem proposed by the user;

[0070] Step S603: Calculate the cosine similarity between the second problem vector and each of the first problem vectors, determine the first problem vector and its corresponding fault handling solution vector when the cosine similarity is greater than the preset similarity threshold, and select multiple fault handling solution vectors from the power grid vector database.

[0071] The second problem vector is determined by identifying the power grid fault problem. A similarity search is performed in the power grid vector database, using cosine similarity as the distance metric. The calculation formula is as follows:

[0072] cos(θ)=(A·B) / (|A|·|B|),

[0073] Where cos(θ) represents the cosine similarity between each first question vector and the second question vector, A represents the second question vector, and B represents each first question vector in the power grid database;

[0074] Based on the calculated cosine similarity results of each first problem vector and second problem vector, the first problem vector and its corresponding fault handling solution vector when the cosine similarity is greater than the preset similarity threshold are determined, thus obtaining multiple fault handling solution vectors.

[0075] In one embodiment, see Figure 7 The diagram illustrates the processing steps of a solution retrieval unit in a power grid fault handling knowledge base management system. This embodiment provides a specific implementation method for retrieving a first power grid fault handling solution from a power grid knowledge graph based on a graph query statement:

[0076] Step S701: Determine each second power grid fault handling scheme corresponding to multiple fault handling scheme vectors selected from the power grid vector database;

[0077] Based on the multiple fault handling scheme vectors obtained by the database filtering module 15, the corresponding second power grid fault handling scheme is determined. The first step of retrieval is performed through the power grid database, and the second step of retrieval is performed through the power grid knowledge graph based on the results of the first step of retrieval.

[0078] Step S703: Determine the content in the power grid knowledge graph that corresponds to each second power grid fault handling plan, and retrieve the first power grid fault handling plan from it based on the graph query statement;

[0079] The first power grid fault handling solution is selected from the power grid knowledge graph by filtering out entities and relationships related to the second power grid fault handling solution. Then, the power grid fault handling solution that best solves the power grid fault problem proposed by the user is selected from the entities and relationships related to the second power grid fault handling solution by using a graph query statement (i.e., a Cypher statement).

[0080] In one embodiment, the power grid fault handling knowledge base management system provided in this embodiment further includes: a training module 16;

[0081] A sample dataset is constructed based on the fault problems of each power grid and their corresponding graph query statements, as well as the fault problems of each power grid and their corresponding power grid fault handling solutions in the power grid knowledge graph.

[0082] The sample dataset is input into the large language model for training to obtain the query retrieval module 13;

[0083] By inputting power grid fault problems and their corresponding graph query statements into the large language model for training, the trained large language model can directly obtain graph query statements (i.e., Cypher statements) from the power grid fault problems proposed by users. By inputting each power grid fault problem and its corresponding power grid fault handling solution from the power grid knowledge graph into the large language model for training, the trained large language model can directly retrieve the first power grid fault handling solution from the power grid knowledge graph through the graph query statement.

[0084] In one embodiment, the power grid fault handling knowledge base management system provided in this embodiment further includes: a performance testing module 17;

[0085] The sample dataset also includes a test dataset;

[0086] The performance testing module 17 is used to input the test dataset into the query retrieval module 13 to obtain performance testing parameters to evaluate the performance of the query retrieval module 13; among which, the performance testing parameters include: precision, recall and F1 score.

[0087] The test dataset is input into the query retrieval module 13 to determine the precision P and recall R. A higher precision can reflect the better performance of the query retrieval module 13 to a certain extent; a higher recall can also reflect the better performance of the query retrieval module 13 to a certain extent.

[0088] Based on precision and recall, the harmonic mean F1 is determined as follows:

[0089]

[0090] The F1 score is the harmonic mean of recall and precision. A high F1 score indicates that the query retrieval module 13 has achieved a balance between reducing false positives (high precision) and reducing false negatives (high recall), resulting in better overall classification performance.

[0091] In one embodiment, the power grid fault handling knowledge base management system provided in this embodiment further includes: a human-computer interaction module 18;

[0092] The human-computer interaction module 18 is used to communicate with the user and to visualize the first power grid fault handling solution retrieved based on the power grid fault problem raised by the user.

[0093] The aforementioned human-computer interaction module 18 provides a chatbot interface, supporting multi-turn question-and-answer sessions. It can also provide retrieved power grid fault handling solutions based on each power grid fault question raised by the user, and visualize these solutions for user review. Specifically, this module adopts a responsive design, adapting to different terminal devices, including PCs, tablets, and mobile devices. The main body of the module is a dialog box, displaying the user-inputted power grid fault question on the left and the retrieved corresponding power grid fault handling solutions on the right. During multi-turn question-and-answer sessions, the module supports rich text formatting, displaying text, tables, images, flowcharts, and other formats. The module's dialogue interface supports multi-turn interaction, remembers context information, and allows users to interact in multiple ways. The module refines questions or requests for more detailed explanations during rounds of dialogue. While providing retrieved power grid fault handling solutions, it also extracts the steps and decisions within these solutions and renders them using the mermaid.js library for clear presentation to the user. If the power grid equipment structure in the fault handling solution is complex, the human-computer interaction module 18 will use 3D model visualization to create a 3D model of the power grid equipment for user review and understanding. This 3D model supports rotation, scaling, and component highlighting. The dialogue interface of the human-computer interaction module 18 also integrates a history session management function, allowing users to view and continue previous conversations, recording frequently used queries, and providing quick access points.

[0094] In one embodiment, the power grid fault handling knowledge base management system provided in this embodiment further includes: an update module 19;

[0095] The update module 19 is used to receive power grid industry-related knowledge data in real time and update the power grid knowledge graph and power grid vector database based on the received power grid industry-related knowledge data in real time.

[0096] The aforementioned update module 19 implements a semi-supervised learning mechanism. When new power grid industry-related knowledge data (i.e., power grid fault problems and their corresponding power grid fault handling solutions) is generated, the new power grid industry-related knowledge data is first initially classified and labeled by the rule engine, and key information (such as fault type, equipment type, handling method, etc.) is extracted. The power grid industry-related knowledge data after initial labeling is verified by the review system to determine whether to retain the power grid industry-related knowledge data. Specifically, the review system is divided into three levels: (1) The primary review is automatically completed by the built-in algorithm in the update module 19, with an accuracy rate of 85%; (2) The intermediate review is pushed by the system to relevant field experts for confirmation; (3) The advanced review is for complex or controversial cases, and experts are organized to discuss and determine the final labeling. The review level is flexibly selected according to the different power grid industry-related knowledge data. The power grid industry-related knowledge data of the cases that have passed the review is added to the data storage module 11. The newly added data in the data storage module 11 updates the power grid vector database. At the same time, the entities and relationships in the newly added data in the data storage module 11 are extracted and added to the power grid knowledge graph to update the power grid knowledge graph.

[0097] The fault handling solution generation system provided in this invention is mainly applicable to power grid operators (such as dispatchers and maintenance personnel), power company operation and maintenance management personnel, smart distribution network system developers and integrators, and emergency response and training personnel. Emergency response and training personnel can use this system for simulation training and supplementary teaching, improving trainees' response capabilities. Smart distribution network system developers and integrators can integrate this system with SCADA, EMS, and other systems to obtain high-performance power intelligent systems. Power grid operators need to build an enterprise-level knowledge base system as an auxiliary decision-making tool; the resulting system possesses high power grid knowledge accumulation, high structure, and high queryability. Power grid operators need to quickly find historical fault cases and emergency handling solutions; they can input the power grid fault problem into the system when a sudden power grid fault occurs to quickly and accurately retrieve the corresponding power grid fault handling solution, improving the efficiency and accuracy of handling power grid fault problems.

[0098] The power grid fault handling knowledge base management system provided in this embodiment of the invention can retrieve power grid fault handling solutions corresponding to power grid fault problems proposed by users based on a trained large language model, providing professional answers that meet the warning requirements and assisting users in making judgments. The power grid vector database supports contextual semantic matching rather than just keyword-based retrieval, enabling rapid retrieval of historical cases semantically similar to the current power grid fault problem, achieving efficient recall of relevant content. In the process of constructing the power grid knowledge graph, the system unifies the expression of power grid industry-related data from different sources (such as dispatch records, operation manuals, expert experience, etc.), achieving the fusion of multi-source heterogeneous knowledge data. Moreover, compared with traditional static knowledge bases, this system has proactive question answering and recommendation functions, and can support the annotation, verification, and automatic integration of new cases into the knowledge base.

[0099] Furthermore, in the description of the embodiments of the present invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in the present invention based on the specific circumstances.

[0100] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0101] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0102] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A power grid fault handling knowledge base management system, characterized in that, include: Data storage module: used to acquire and store knowledge data related to the power grid industry; wherein, the knowledge data related to the power grid industry includes power grid fault problems and corresponding power grid fault handling solutions; Knowledge graph construction module: used to construct a power grid knowledge graph based on the power grid industry-related knowledge data; The query and retrieval module includes a statement generation unit and a solution retrieval unit. The statement generation unit generates a graph query statement based on the power grid fault problem proposed by the user. The solution retrieval unit retrieves a first power grid fault handling solution from the power grid knowledge graph based on the graph query statement. The query and retrieval module is trained using a training dataset. The training dataset includes power grid fault problems labeled with graph query statements and a power grid knowledge graph.

2. The power grid fault handling knowledge base management system according to claim 1, characterized in that, The construction of the power grid knowledge graph based on the power grid industry-related knowledge data includes: Determine the entities and relationships between them in the power grid fault problem and its corresponding power grid fault handling plan; Based on the entities and the relationships between them, triples are constructed, and a power grid knowledge graph is built based on the triples.

3. The power grid fault handling knowledge base management system according to claim 1, characterized in that, Also includes: Database building module and database filtering module; The database construction module is used to input the power grid industry-related knowledge data into the natural language processing module for processing, and generate a power grid vector database. The database filtering module is used to filter multiple fault handling solution vectors from the power grid vector database based on the power grid fault problem proposed by the user.

4. The power grid fault handling knowledge base management system according to claim 3, characterized in that, The step of inputting the power grid industry-related knowledge data into a natural language processing module for processing to generate a power grid vector database includes: Based on the natural language processing module, the power grid fault problem and its corresponding power grid fault handling solution are processed by text segmentation to obtain multiple problem text blocks and their corresponding fault handling solution text blocks. Each of the aforementioned problem text blocks and its corresponding fault handling solution text blocks are converted into a first problem vector and its corresponding fault handling solution vector, respectively, to obtain a power grid vector database.

5. The power grid fault handling knowledge base management system according to claim 4, characterized in that, The process involves selecting multiple fault handling solution vectors from the power grid vector database based on the power grid fault problem proposed by the user, including: Determine the corresponding second problem vector based on the power grid fault problem proposed by the user; Calculate the cosine similarity between the second problem vector and each of the first problem vectors, determine the first problem vector and its corresponding fault handling solution vector when the cosine similarity is greater than a preset similarity threshold, and select multiple fault handling solution vectors from the power grid vector database.

6. The power grid fault handling knowledge base management system according to claim 5, characterized in that, The step of retrieving a first power grid fault handling solution from the power grid knowledge graph based on the graph query statement includes: Determine each second power grid fault handling scheme corresponding to multiple fault handling scheme vectors selected from the power grid vector database; The content corresponding to each of the second power grid fault handling schemes in the power grid knowledge graph is determined, and the first power grid fault handling scheme is retrieved from it based on the graph query statement.

7. The power grid fault handling knowledge base management system according to claim 1, characterized in that, Also includes: Training module; A sample dataset is constructed based on the fault problems of each power grid and their corresponding graph query statements, as well as the fault problems of each power grid and their corresponding power grid fault handling solutions in the power grid knowledge graph. The sample dataset is input into a large language model for training to obtain the query retrieval module.

8. The power grid fault handling knowledge base management system according to claim 7, characterized in that, Also includes: Performance testing module; The sample dataset also includes a test dataset; The performance testing module is used to input the test dataset into the query retrieval module to obtain performance testing parameters to evaluate the performance of the query retrieval module; wherein, the performance testing parameters include: precision, recall and F1 score.

9. The power grid fault handling knowledge base management system according to claim 1, characterized in that, Also includes: Human-computer interaction module; The human-computer interaction module is used to communicate with the user and to visually display the first power grid fault handling solution retrieved based on the power grid fault problem raised by the user.

10. The power grid fault handling knowledge base management system according to claim 1, characterized in that, Also includes: Update module; The update module is used to receive power grid industry-related knowledge data in real time, and update the power grid knowledge graph and power grid vector database based on the received power grid industry-related knowledge data in real time.

Images