Power generation field equipment question and answer method and system based on knowledge graph

Abstract

The invention relates to the field of power generation and the technical field of artificial intelligence, in particular to a power generation field equipment question answering method and system based on a knowledge graph, and the method comprises the steps: obtaining a power generation field equipment knowledge question inputted by a user; inputting the user question into a semantic analysis model, and converting the input question into a knowledge graph query statement by the semantic analysis model; performing retrieval by utilizing the knowledge graph in the power generation field to correspondingly obtain a retrieval result; generating a corresponding question answer according to the retrieval result; and the knowledge graph of the power generation field is constructed according to multi-source heterogeneous data. According to the method, through the synergistic effect of the semantic analysis model and the knowledge graph, intelligent conversion from user questions to knowledge answers is achieved, and the accuracy and efficiency of equipment knowledge questions and answers in the power generation field are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of power generation and the field of artificial intelligence technology, and particularly relates to a method and system for equipment question answering in the power generation field based on a knowledge graph. Background Art

[0002] Currently, the mainstream large language models (LLMs) are based on the Transformer architecture and rely on massive multi-source corpora for training, demonstrating powerful natural language understanding and generation capabilities. Thanks to the wide coverage of training data, the question answering systems based on LLMs perform well in general fields. However, to build a question answering system in the power generation field, in-depth professional knowledge in the vertical field is required. Relying solely on the mainstream LLMs cannot provide professional answers, and the training of LLMs depends on static training data, making it difficult to adapt to dynamic industry needs. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a method and system for equipment question answering in the power generation field based on a knowledge graph to solve the technical problem that the reasoning ability of the current knowledge question answering system in the power generation field is poor, resulting in inaccurate answers.

[0004] The object of the present invention is achieved by the following technical solutions: In a first aspect, the present invention provides a method for equipment question answering in the power generation field based on a knowledge graph, including: Obtaining a knowledge question about equipment in the power generation field input by a user; Inputting the user's question into a semantic parsing model, which converts the user's question into a knowledge graph query statement; retrieving using the knowledge graph in the power generation field to obtain corresponding retrieval results; generating corresponding question answers according to the retrieval results; the semantic parsing model uses a large language model; The knowledge graph in the power generation field is constructed based on multi-source heterogeneous data.

[0005] As a further improvement of the present invention, the construction method of the knowledge graph in the power generation field is: Obtaining multi-source heterogeneous data in the power generation field; Obtaining unstructured data and structured data from the multi-source heterogeneous data, and respectively using different methods to construct nodes, attributes, and link relationships of the knowledge graph according to the structured data and unstructured data; Performing conflict resolution and redundancy merging on the obtained unstructured data and structured data to generate a unified entity identifier; Performing data fusion according to the entity identifier; storing the fused data in a Neo4j graph database to form a knowledge graph in the power generation field.

[0006] As a further improvement of the present invention, the structural data processing method includes: Regarding the data table as a node of the knowledge graph; regarding the data relationship between tables as an edge, and mapping the content of table fields to node attributes; Regarding the table content as reference data, and enhancing the relationship density between entities through the association rules between table fields.

[0007] As a further improvement of the present invention, the unstructured data processing method specifically includes: Using named entity recognition to extract triples from unstructured data, where the triples include device names, fault types, and policy clause entities; Based on a pre-trained relation extraction model, establishing triples regarding device names, fault types, and policy clause entities, and obtaining a triple inference link regarding fault-device-solution; The relation extraction model is the BERT-CRF model.

[0008] As a further improvement of the present invention, the semantic parsing model converts the input question into a knowledge graph query statement, specifically including: The semantic parsing model utilizes the structured information of knowledge graph elements, and through a preset prompt template, converts the input question into the query language cypher of the knowledge graph.

[0009] As a further improvement of the present invention, it further includes a question-answering assistant, which generates corresponding question answers according to the retrieval results, specifically including: The question-answering assistant is composed of a large language model, and uses the Qwen2.5-32B-Instruct model as the inference model of the question-answering assistant. Input the retrieval result and the user's question into the question-answering inference model, and correspondingly output the inference process and the final inference result.

[0010] As a further improvement of the present invention, a timestamp attribute is further added to the device nodes in the power generation field knowledge graph to form a time-series subgraph, specifically including: When performing knowledge graph retrieval, based on the timestamp attribute of the device node parameters, constructing a device status time-series chain; Using the time-series subgraph network to infer the potential impact path of historical events on the current fault, and predicting the event occurrence probability.

[0011] In a second aspect, the present invention provides a power generation field device question-answering system based on a knowledge graph, which is used to implement the above-mentioned power generation field device question-answering method based on a knowledge graph, including: A data acquisition module, which is used to acquire the knowledge questions of power generation field devices input by the user; Knowledge Q&A module, input the user's question into the semantic parsing model, and the semantic parsing model converts the user's question into a knowledge graph query statement; retrieve using the knowledge graph in the power generation field to obtain the corresponding retrieval result; generate the answer to the corresponding question according to the retrieval result. Knowledge graph construction module, the knowledge graph in the power generation field is constructed from multi-source heterogeneous data.

[0012] In a third aspect, the present invention provides a computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to execute the above-mentioned power generation field device Q&A method based on a knowledge graph.

[0013] In a fourth aspect, the present invention provides a computing device, including: One or more processors, a memory, and one or more programs, where the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include steps for executing the above-mentioned power generation field device Q&A method based on a knowledge graph.

[0014] The beneficial effects of the present invention are as follows: The present invention provides a power generation field device Q&A method based on a knowledge graph. By using a large language model as the semantic parsing model, the system can accurately convert the natural language questions input by users into structured knowledge graph query statements, which not only improves the accuracy of queries but also can handle complex natural language expressions. Retrieving using the knowledge graph in the power generation field can quickly obtain accurate information related to the user's question. The structured characteristics of the knowledge graph make information retrieval more efficient and accurate. Based on the retrieval results, the system can generate accurate and detailed answers. The present invention not only improves the accuracy of the answers but also can provide customized information according to the specific needs of the user's question.

[0015] Furthermore, by obtaining multi-source heterogeneous data in the power generation field, the system can integrate data from different channels and different formats, and process unstructured data and structured data separately. For unstructured data, natural language processing techniques are used to extract key information; for structured data, its structured characteristics are directly utilized to construct nodes, attributes, and relationships. This method ensures the comprehensive utilization of data and avoids information omission. During the data fusion process, by resolving conflicts and merging redundancies between unstructured data and structured data, a unified entity identifier is generated. This ensures the uniqueness and consistency of the data in the knowledge graph and avoids problems caused by data redundancy and conflicts.

[0016] Furthermore, the semantic parsing model utilizes the structured information of knowledge graph elements (such as nodes, attributes, relationships, etc.). Through a preset prompt template, it accurately converts the user's question into the query language of the knowledge graph, thus ensuring the accuracy and relevance of the query and avoiding information retrieval errors caused by language ambiguity or vagueness. By combining the structured information of the knowledge graph, the semantic parsing model can better understand the context and intention of the user's question, and thus generate query statements that better meet the user's needs. This method makes full use of the advantages of the knowledge graph and improves the comprehensiveness and depth of information retrieval. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention 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 some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 is a flowchart of a method for answering questions about power generation equipment based on a knowledge graph in an embodiment of the present invention; Figure 2 is a schematic diagram of knowledge graph construction in an embodiment of the present invention; Figure 3 is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to make the purpose and technical solutions of the present invention clearer and easier to understand. The following further details the present invention in conjunction with the drawings and embodiments. The specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0020] The technical solutions of the present invention will be clearly and completely described below in conjunction with the drawings and specific embodiments, where the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0021] Embodiment 1 As Figure 1 and Figure 2 shown, this embodiment provides a method for answering questions about power generation equipment based on a knowledge graph. The following are the specific implementation manners.

[0022] First, obtain the user's input knowledge question about power generation equipment.

[0023] Then, the user's question is input into the semantic parsing model, which converts the user's question into a knowledge graph query statement; the knowledge graph of the power generation field is used for retrieval, and the corresponding retrieval results are obtained; corresponding question answers are generated according to the retrieval results; the semantic parsing model adopts a large language model. For example, the semantic parsing model in this embodiment adopts the LLM model.

[0024] Among them, the knowledge graph of the power generation field is constructed based on multi-source heterogeneous data. The multi-source heterogeneous data in this embodiment includes relevant industry standards, specification documents obtained from the power plant SIS system, equipment manufacturer APIs, and the national standard file library, and historical power generation field equipment maintenance data is collected, including fault phenomena, maintenance measures, maintenance time, relevant field articles, and literature materials.

[0025] In this embodiment, the construction method of the power generation field knowledge graph is as follows: Obtain multi-source heterogeneous data in the power generation field; obtain unstructured data and structured data from the multi-source heterogeneous data, and respectively adopt different methods to construct the nodes, attributes, and link relationships of the knowledge graph according to the structured data and unstructured data; perform conflict resolution and redundancy merging on the obtained unstructured data and structured data to generate a unified entity identifier; perform data fusion according to the entity identifier; store the fused data in the Neo4j graph database to form a knowledge graph of the power generation field. Data from different sources may have conflicts and redundancies, resulting in contradictions or overlaps in the generated knowledge graph. Knowledge fusion aligns the conflicting parts and merges the overlapping parts to ensure that each piece of information in the knowledge graph is accurate and unique, avoiding confusion and errors.

[0026] The structural data in this embodiment mainly includes the production information and equipment ledger of power plant enterprises, which are stored in the form of the current state of the table. The unstructured data mainly includes accident reports and maintenance records of power plant equipment, which are mainly stored in the form of text or pictures.

[0027] Among them, the processing method of structured data includes: Use the data table as a node of the knowledge graph; the data relationship between tables is used as an edge, and the table field content is mapped to the node attribute; use the table content as reference data to increase the density of the relationship between entities through the association rules between table fields. Taking the equipment operation parameter table in the power plant SIS system as an example, the table name is used as the node name, and fields such as equipment ID, parameter name, and value are used as node attributes, and the relationship with other tables is established through association rules (such as equipment ID matching) to increase the density of the relationship between entities.

[0028] The processing method of unstructured data specifically includes: Extract triples from unstructured data using named entity recognition. The triples include device names, fault types, and policy clause entities. Based on a pre-trained relation extraction model, establish triples regarding device names, fault types, and policy clause entities to obtain multiple inference links for the fault-device-solution triples.

[0029] The relation extraction model in this embodiment is the BERT-CRF model.

[0030] To enable the knowledge graph to accurately recognize the current natural language, the semantic parsing model utilizes the structured information of the knowledge graph elements and maps the input question to the query language cypher of the knowledge graph through a preset prompt template. For example, the prompt template is: "You are an expert in the power generation field. Answer the question based on the following knowledge: {retrieved device parameters, fault solutions, policy clauses} Question: {user input}".

[0031] In addition, a timestamp attribute is added to the device nodes in the power generation field knowledge graph to form a temporal subgraph, specifically including: When performing knowledge graph retrieval, based on the timestamp attribute of the device node parameters, construct a device status temporal chain; use the temporal subgraph network to infer the potential impact path of historical events on the current fault and predict the event occurrence probability. For example, for the operating parameters of a certain device, such as temperature, pressure, etc., connect them in chronological order to form a temporal chain reflecting the change of the device status over time. By analyzing a series of events (such as device maintenance time, ambient temperature change, etc.) before the device had similar faults in the past and their time relationship with the current fault, predict the event occurrence probability. For example, through analysis by the temporal subgraph network, it is found that the ambient temperature has continued to rise within one week before the device overheats each time in the past, thus predicting a higher probability of the device overheating when the current ambient temperature rises.

[0032] The question answering assistant uses the Qwen2.5-32B-Instruct model as the inference model, inputs the retrieval results and the user's question into the large language model, and in the large language model, the inference process and the final inference result are output correspondingly.

[0033] In addition, the knowledge graph in the power generation field of this embodiment is updated at a set period. Specifically, it includes updating real-time data such as policy releases and device changes.

[0034] Embodiment 2 As a preferred implementation manner in Embodiment 1, this embodiment specifically includes: Obtain raw data from the power plant SIS system, device manufacturer APIs, and national standard document libraries, and clean and standardize the obtained data.

[0035] Collect and organize messy multi-source data, and use various methods to process different types of data. For structured data such as production information and equipment ledgers in power plant enterprises, the storage fields are generally complex and diverse, and there are correlations between some data tables. When constructing a knowledge graph, the data tables are used as the nodes of the knowledge graph, the relevant data as the relationships, and the table content as the attributes to form the initial structure of the knowledge graph. Regarding the field information of the data tables, the table content is used as a reference to extract the relationships between different fields, improving the entity association density of the knowledge graph. For unstructured data such as accident reports and maintenance records of power plant equipment, most of them are stored in text or picture form. How to extract triples from the data is the difficulty in constructing the knowledge graph. Therefore, natural language processing technology is introduced, and named entity recognition and relation extraction models are used to extract triples in the data to build the link of the knowledge graph.

[0036] Specifically, for structured data, the table content is used as node information, and the foreign keys of the table data are used to construct the relationships between the tables; for unstructured data, the BERT-CRF model is used to extract triples in the data to construct the relationships between the subjects. In the knowledge fusion stage, all the nodes, attributes, and relationships of the data are integrated to eliminate conflicts and redundancies in the knowledge graph. The constructed knowledge graph is stored in the Neo4j graph database, with a one-month update cycle set. Update data is regularly obtained from the data source, and batch updates of the knowledge graph are performed. Combined with the expert opinions in the power generation field, the updated part is reviewed to ensure the accuracy and consistency of the knowledge graph.

[0037] Use the LLM to convert the user's question into a graph query statement, and use the prompt technology to specify the generation task of the LLM, enabling the model to focus on the entity information in the user input to obtain the name of the query node. In addition, the output format of the model is restricted to the query language cypher to obtain the graph query statement corresponding to the user's question. For example, the designed prompt template is: "You are an expert in the power generation field. Answer the question based on the following knowledge: {retrieved equipment parameters, fault solutions, policy terms} Question: {user input}".

[0038] Traverse the Neo4j graph database according to the query language cypher to obtain the retrieval results. At the same time, introduce a professional vocabulary library in the power generation field to query node information with the same semantics, and summarize and integrate the retrieval results as the context information of the LLM.

[0039] Deploy the Qwen2.5-32B-Instruct model as the inference model of the question-answering assistant, input the retrieval results and the user's query into the Qwen2.5-32B-Instruct model, combine graph retrieval and large model generation to avoid "hallucinatory answers" and improve the credibility of the answers.

[0040] Transform multi-source heterogeneous data in the power generation industry into inferable structured knowledge through a knowledge graph, combine it with mainstream large language models (LLMs), and build a question-answering assistant in the vertical field of the power generation industry, improving the efficiency of industry knowledge management and the level of intelligence.

[0041] This embodiment is also described in detail with specific examples. The specific steps are as follows: First, the user inputs: "How to calculate the carbon emission quota of a coal-fired power plant?". The semantic parsing module performs semantic parsing: matches the system node named "coal-fired power plant" and its associated nodes. The knowledge graph retrieves: retrieves the node file and extracts the carbon emission quota calculation formula. The large language model generates an answer: "Quota = benchmark value × actual output × adjustment coefficient. See Article × of the 'Implementation Plan for the Carbon Market' for details."

[0042] Embodiment 3 This embodiment provides a device question-answering system in the power generation field based on a knowledge graph, which is used to implement the device question-answering method in the power generation field based on the knowledge graph in Embodiment 1 and Embodiment 2. This system specifically includes: A data acquisition module, which is used to acquire the knowledge questions of power generation field devices input by the user; A knowledge question-answering module, which inputs the user's question into the semantic parsing model. The semantic parsing model converts the user's question into a knowledge graph query statement; retrieves using the knowledge graph in the power generation field and obtains the corresponding retrieval result; generates an answer to the corresponding question according to the retrieval result; A knowledge graph construction module, and the knowledge graph in the power generation field is constructed based on multi-source heterogeneous data. The knowledge graph construction module specifically includes: acquiring multi-source heterogeneous data in the power generation field; acquiring unstructured data and structured data from the multi-source heterogeneous data, and respectively using different methods to construct the nodes, attributes, and link relationships of the knowledge graph according to the structured data and unstructured data; performing conflict resolution and redundant merging on the acquired unstructured data and structured data to generate a unified entity identifier; performing data fusion according to the entity identifier; storing the fused data in the Neo4j graph database to form a knowledge graph in the power generation field.

[0043] Among them, the processing method for structured data includes: taking data tables as nodes of the knowledge graph; taking the data relationships between tables as edges, and mapping the table field contents to node attributes; taking the table contents as reference data and enhancing the relationship density between entities through the association rules between table fields.

[0044] Rather than structured data processing methods, specifically including: using named entity recognition to extract triples from unstructured data, where the triples include device names, fault types, and policy clause entities; establishing triples regarding device names, fault types, and policy clause entities based on a pre-trained relation extraction model to obtain a triple inference link for fault-device-solution; where the relation extraction model is a BERT-CRF model.

[0045] Embodiment 4 In an embodiment of the present invention, a computer-readable storage medium is provided. This medium belongs to the memory device of the terminal device and is mainly used to store programs and data. The computer-readable storage medium includes both the storage medium built into the terminal and the extended storage medium supported by the terminal. Specifically, any tangible medium that can store a program and be used by an instruction execution system, device, or component belongs to this category. This storage medium provides storage space for storing the terminal operating system and instructions that can be loaded and executed by the processor (including one or more computer programs and their code). Examples include electrically connected devices, portable disks, hard disks, RAM, ROM, EPROM / flash memory, optical fibers, CD-ROMs, optical storage devices, magnetic storage devices, etc., and combinations thereof.

[0046] In addition, the computer-readable storage medium also relates to data signals propagated in the baseband or as a carrier wave. These signals carry readable program code and can be in the form of electromagnetic signals, optical signals, etc. The readable storage medium is not limited to the above types and also includes other media that can send, propagate, or transmit a program for use by an instruction execution system, device, or component. The program code can be transmitted via wireless, wired, optical fiber, RF, etc.

[0047] The program code can be written in various programming languages, such as object-oriented languages (Python, Java, C++, etc.) and procedural languages (C language, etc.). The code can be executed completely or partially on the user device, or can be used as an independent software package, or be executed partially / fully on a remote device. The remote device is connected to the user device via a LAN, WAN, or Internet service provider.

[0048] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the method and system for device Q&A in the power generation field based on the knowledge graph in the above embodiments; Obtain a knowledge question about a power generation field device input by the user; Input the input question into the semantic parsing model. The semantic parsing model converts the input question into a knowledge graph query statement; use the knowledge graph of the power generation field for retrieval to obtain the corresponding retrieval result; generate the corresponding question answer according to the retrieval result; the semantic parsing model uses a large language model; The knowledge graph in the field of power generation is constructed based on multi-source heterogeneous data.

[0049] Embodiment 5 Figure 3 This is a block diagram of an electronic device provided by the present invention according to an embodiment.

[0050] Please refer to Figure 3 , the terminal device 600 is an electronic device, and the electronic device is presented in the form of a general computing device. The components of the electronic device may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.

[0051] Among them, the storage unit stores program codes, and the program codes can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the method part of this specification. For example, the processing unit 610 can execute the steps as shown in Figure 1 .

[0052] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 6201 and / or a cache storage unit 6202, and may further include a read-only storage unit (ROM) 6203.

[0053] The storage unit 620 may further include a program / utility 6204 having a set (at least one) of program modules 6205. Such program modules 6205 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. The implementation of a network environment may be included in each or some combination of these examples.

[0054] The bus 630 may represent one or more of several types of bus structures, including a storage unit bus or a storage unit controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any one of the multiple bus structures.

[0055] The electronic device 600 can also communicate with one or more external devices 700 (such as a keyboard, a pointing device, a Bluetooth device, etc.), and can also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (such as a router, a modem, etc.). Such communication can be carried out through the input / output (I / O) interface 650. Moreover, the electronic device 600 can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through the network adapter 660. The network adapter 660 can communicate with other modules of the electronic device 600 through the bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

Claims

1. A method for equipment question answering in the power generation field based on a knowledge graph, characterized in that Including: Obtain equipment knowledge questions in the power generation field input by the user; Input the input question into the semantic parsing model, and the semantic parsing model converts the input question into a knowledge graph query statement; Retrieve using the knowledge graph in the power generation field, and obtain the corresponding retrieval results; generate corresponding question answers according to the retrieval results; the semantic parsing model uses a large language model; The knowledge graph in the power generation field is constructed based on multi-source heterogeneous data.

2. The method for answering questions about power generation field equipment based on a knowledge graph according to claim 1, wherein The construction method of the knowledge graph in the power generation field is as follows: Obtain multi-source heterogeneous data in the power generation field; Obtain unstructured data and structured data from the multi-source heterogeneous data, and respectively adopt different methods to construct the nodes, attributes and link relationships of the knowledge graph according to the structured data and unstructured data; Perform conflict resolution and redundancy merging on the obtained unstructured data and structured data to generate a unified entity identifier; Perform data fusion according to the entity identifier; store the fused data in the Neo4j graph database to form a knowledge graph in the power generation field.

3. The method for answering questions about power generation field equipment based on a knowledge graph according to claim 2, wherein The structured data processing method includes: Use the data table as the node of the knowledge graph; the data relationship between the tables as the edge, and map the table field content to the node attribute; Use the table content as reference data, and improve the relationship density between entities through the association rules between table fields.

4. The method for equipment question answering in the power generation field based on the knowledge graph according to claim 2, wherein The unstructured data processing method specifically includes: Adopt named entity recognition to extract triples in the unstructured data, and the triples include equipment name, fault type, and policy clause entities; Based on a pre-trained relation extraction model, establish triples about equipment name, fault type, and policy clause entities to obtain a triple inference link about fault-equipment-solution; The relation extraction model is a BERT-CRF model.

5. The method for equipment question-answering in the power generation field based on a knowledge graph according to claim 1, wherein The semantic parsing model converts the input question into a knowledge graph query statement, specifically including: The semantic parsing model uses the structured information of the knowledge graph elements, and through a preset prompt template, converts the input question into the query language cypher of the knowledge graph.

6. The method for answering questions about equipment in the power generation field based on a knowledge graph according to claim 1, characterized in that It also includes a question and answer assistant, and the question and answer assistant generates corresponding question answers according to the retrieval results, specifically including: The question and answer assistant is composed of a large language model, and uses the Qwen2.5-32B-Instruct model as the inference model of the question and answer assistant. Input the retrieval results and the user's question into the large language model, and the large language model outputs the inference process and the final inference result.

7. The method for answering questions about power generation field equipment based on a knowledge graph according to claim 2, wherein, In the knowledge graph in the power generation field, a timestamp attribute is also added to the equipment node to form a time series subgraph, specifically including: When performing knowledge graph retrieval, construct an equipment status time series chain based on the timestamp attribute of the equipment node parameters; Adopt a time series subgraph network to infer the potential impact path of historical events on the current fault, and predict the event occurrence probability.

8. A device Q&A system in the power generation field based on a knowledge graph, which is used to implement the method for device Q&A in the power generation field based on a knowledge graph according to any one of claims 1 to 7, characterized in that, Including: A data acquisition module for obtaining equipment knowledge questions in the power generation field input by the user; A knowledge question and answer module that inputs the input question into the semantic parsing model, and the semantic parsing model converts the input question into a knowledge graph query statement; Retrieve using the knowledge graph in the power generation field, and obtain the corresponding retrieval results; generate corresponding question answers according to the retrieval results; Knowledge graph construction module, and the knowledge graph in the power generation field is constructed based on multi-source heterogeneous data.

9. A computer-readable storage medium storing one or more programs, characterized in that, The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform the knowledge graph-based power generation field device Q&A method according to any one of claims 1 to 7.

10. A computing device, characterized in that, Including: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include steps for performing the knowledge graph-based power generation field device Q&A method according to any one of claims 1 to 7.

Images