Reflexion and knowledge graph-based power operation and maintenance knowledge capability measurement method, system and equipment

Abstract

The invention relates to the technical field of operation and maintenance of power grid main equipment, in particular to a method, a system and equipment for measuring power operation and maintenance knowledge capability based on Reflexion and a knowledge graph, and the method comprises the following steps: identifying and extracting an operation and maintenance entity in any question and answer pair, and according to the operation and maintenance entity, determining the operation and maintenance knowledge capability of the power grid main equipment; querying associated information corresponding to the electric power operation and maintenance map in the electric power operation and maintenance map; adding the associated information into a predefined chained thinking prompt template, and according to the chained thinking prompt template, generating a reasoning process and an answer of a next question answering; and if the current question answering times do not exceed the set question answering limit times, scoring the answers to obtain evaluation scores of the answers, and performing reflection analysis according to the evaluation scores to obtain final answers of the answers. According to the method, the measurement requirement of the electric power operation and maintenance knowledge in the large language model can be met, and certain help is provided for the electric power large language model knowledge question-answer task and construction work in the subsequent specific business background.

Description

Technical Field

[0001] The present invention relates to the technical field of operation, maintenance and repair of main equipment of power grids, and in particular to a method, system and equipment for measuring knowledge capability of power operation, maintenance and repair based on Reflexion and knowledge graph. Background Art

[0002] The operation, maintenance and overhaul of the main equipment of the power grid constitute an indispensable key link in the power system, and its core focuses on the maintenance and overhaul of core components such as generators, transformers, and switchgear. With the continuous increase in electricity demand and the profound adjustment of the energy structure, the operation and maintenance personnel are increasingly eager for knowledge related to the main equipment to ensure that they can complete equipment maintenance and fault handling tasks efficiently and accurately.

[0003] The traditional operation and maintenance method mainly relies on manual experience and on-site judgment. This method is not only inefficient, but also easily affected by human factors, resulting in the quality and safety of operation and maintenance cannot be fully guaranteed.

[0004] In order to overcome the shortcomings of traditional operation and maintenance methods, knowledge graphs, as a structured knowledge representation method, can clearly display operation and maintenance entities and their relationships, providing strong support for intelligent reasoning and decision-making. Artificial intelligence technology, especially natural language processing and machine learning technology, can achieve automatic understanding and answering of operation and maintenance problems, greatly improving the efficiency and accuracy of operation and maintenance.

[0005] However, the existing methods for measuring the knowledge and ability of power operation and maintenance based on knowledge graphs and artificial intelligence technologies still have some shortcomings. For example, these methods can only handle simple question-answer pairs when identifying operation and maintenance entities and related information. For complex operation and maintenance scenarios and problems, their recognition and reasoning capabilities need to be improved. In addition, these methods often lack sufficient reflection and iteration mechanisms when generating answers, resulting in the generated answers being inaccurate and incomplete. Summary of the invention

[0006] In order to solve the above technical problems, the present invention provides a method for measuring the knowledge ability of power operation and maintenance based on Reflexion and knowledge graph, comprising the following steps:

[0007] S1: Identify and extract the operation and maintenance entity in any question-answer pair, and query the corresponding associated information in the power operation and maintenance map according to the operation and maintenance entity;

[0008] S2: adding the associated information to a predefined chain-of-thought prompt template, and generating a reasoning process and an answer for the next question according to the chain-of-thought prompt template;

[0009] S3: Determine whether the number of times the current question is answered exceeds the set answer limit:

[0010] If not, score the answer to obtain an evaluation score for the answer, and reflect and analyze based on the evaluation score to obtain a final answer to the question;

[0011] If so, the answer is the final answer to the question.

[0012] In one embodiment of the present invention, in S2, the associated information is added to a predefined chained thinking prompt template. The expression in can be expressed as:

[0013]

[0014] Among them, t+1 represents the number of rounds of the next round of answering process of the agent; p I is the question-answer instruction, p E For the sample answer, p Q For questions and options, p K represents the associated information, represents the updated reasoning process to be used in the t+1th round of answering, Represents the updated reflection text that will be used in the t+1 round of answering.

[0015] In one embodiment of the present invention, in S3, the calculation expression for obtaining the evaluation score score of the answer is as follows:

[0016]

[0017] Where a is the answer generated according to the chain thinking prompt template, a * is the candidate answer in the question-answer pair, len(x) represents the function used to obtain the length of string x, and the value range of score is between [0,1].

[0018] In one embodiment of the present invention, in S3, the method of performing reflective analysis based on the evaluation score to obtain the final answer to the question is as follows:

[0019] Determine whether the evaluation score reaches a set first threshold:

[0020] If not, reflect and analyze the answer to generate a reflection text, update the chain thinking prompt template according to the reasoning process and the reflection text, obtain an updated chain thinking prompt template, return to step S2, and generate the reasoning process and answer for the next question according to the updated chain thinking prompt template;

[0021] If so, it means that the answer completely matches the answer in the question-answer pair, which is the final answer to the question.

[0022] In one embodiment of the present invention, the expression for updating the chained thinking prompt template according to the reasoning process and the reflection text is as follows:

[0023]

[0024] in, represents the updated reasoning process to be used in the t+1th round of answering, represents the reasoning process generated during the t-th round of answering; represents the updated reflection text to be used in the t+1 round of answers, Represents the reflection text generated during the tth round of answers.

[0025] In one embodiment of the present invention, the reflection text includes three parts: problems that appear in the answer, causes of the problems and corresponding solutions. Among them, the problems that appear in the answer include omissions and wrong selections. According to the problems that appear in the answer, the current problem is analyzed, the causes of the problems are obtained, and corresponding solutions are formulated.

[0026] In one embodiment of the present invention, the power operation and maintenance graph is a directed graph with operation and maintenance entities as nodes and the relationships between operation and maintenance entities as edges; wherein the operation and maintenance entities include power equipment information, fault information and maintenance methods, and a probability value is assigned to each edge, and the probability value serves as the confidence level of the existence of the connection relationship between the operation and maintenance entities.

[0027] In one embodiment of the present invention, the question-answer pair is a triple data structure consisting of a question, an option set, and an answer set.<Q,O,A> ; Where Q represents the power operation and maintenance problem raised by the user, O = {O1, O2, ..., O n} represents a set of options, n≥1, each option O i is a candidate answer to question Q, Represents the answer set, which is a collection of one or more options in option set O that are determined to be the correct answer.

[0028] Based on the same inventive concept, the present invention also provides a power operation and maintenance knowledge capability measurement system based on Reflexion and knowledge graph, which is used to implement the steps of the power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph, and the system includes the following modules:

[0029] The operation and maintenance entity identification and associated information query module is used to identify and extract the operation and maintenance entity in any question-answer pair, and query the associated information corresponding to the operation and maintenance entity in the power operation and maintenance map according to the operation and maintenance entity;

[0030] A chain thinking prompt template construction and answer generation module, used to add the associated information to a predefined chain thinking prompt template, and generate a reasoning process and answer for the next question according to the chain thinking prompt template;

[0031] The answer scoring and final answer determination module is used to score the answer if the number of current answers to the question does not exceed the set answer limit, obtain the evaluation score of the answer, and reflect and analyze based on the evaluation score to obtain the final answer to the question.

[0032] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the method for measuring the knowledge capability of electric power operation and maintenance based on Reflexion and knowledge graph are implemented.

[0033] The above technical solution of the present invention has the following advantages compared with the prior art:

[0034] The present invention takes the generative pre-trained language model as the core, designs a knowledge graph module, extracts operation and maintenance entities in questions and options, retrieves relevant entities and corresponding description information in the domain graph as the question context, and uses the Reflexion framework to answer and reflect on questions in the field of operation and maintenance multiple times, thereby giving full play to the knowledge capabilities of the large language model agent in the field of operation and maintenance, so as to better complete the knowledge capability measurement work.

[0035] In addition, the invention opens up a new perspective and path for the measurement of power operation and maintenance knowledge in large language models. It can not only effectively evaluate the knowledge mastery of the model in this professional field, but also promote business personnel to have a deeper understanding and more reasonable application of the knowledge level of the operation and maintenance knowledge question-answering model. This not only helps to improve the efficiency and accuracy of operation and maintenance work, but also injects new vitality into the intelligent transformation and sustainable development of the power industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0037] Figure 1It is a flow chart of the method for measuring the knowledge capability of electric power operation and maintenance based on Reflexion and knowledge graph provided in the first embodiment of the present invention;

[0038] Figure 2 It is a specific flow chart of the method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph provided in the first embodiment of the present invention;

[0039] Figure 3 is the ratio of completely correct and partially correct results of the method of the present invention when only the Reflexion framework is used;

[0040] Figure 4 is the ratio of completely correct and partially correct results of the method of the present invention when using the Reflexion framework and the knowledge graph;

[0041] Figure 5 It is a schematic diagram of the system structure of the electric power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph provided in the second embodiment of the present invention;

[0042] Explanation of the accompanying drawings in the specification: 100, operation and maintenance entity recognition and related information query module; 200, chain thinking prompt template construction and answer generation module; 300, answer scoring and final answer determination module. DETAILED DESCRIPTION

[0043] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0044] Embodiment 1

[0045] Reference Figure 1 and Figure 2 As shown, the present invention provides a method for measuring power operation and maintenance knowledge capability based on Reflexion and knowledge graph, comprising the following steps:

[0046] S1: Identify and extract the operation and maintenance entity in any question-answer pair, and query the corresponding associated information in the power operation and maintenance map according to the operation and maintenance entity;

[0047] S2: adding the associated information to a predefined chain thinking prompt template, and generating a reasoning process and an answer for the next question according to the chain thinking prompt template;

[0048] S3: Determine whether the number of times the current question is answered exceeds the set answer limit:

[0049] If not, the answer is scored to obtain an evaluation score of the answer, and a reflective analysis is performed based on the evaluation score under the Reflexion framework to obtain a final answer to the question;

[0050] If so, the answer is the final answer to the question.

[0051] Specifically, the question-answer pair is a triple data structure consisting of a question, an option set, and an answer set.<Q,O,A> ; Where Q represents the power operation and maintenance problem raised by the user, O = {O1, O2, ..., O n} represents a set of options, n≥1, each option O i is a candidate answer to question Q, Represents the answer set, which is a collection of one or more options in option set O that are determined to be the correct answer.

[0052] In step S1, identify and extract the operation and maintenance entities in any question-answer pair. If a sentence contains multiple operation and maintenance entities, all of them need to be identified and listed. The returned results must be in the following JSON format: {"entity":["entity 1","entity 2",...]}; if no power entity is identified, please return an empty entity list.

[0053] The power operation and maintenance graph is a directed graph with operation and maintenance entities as nodes and the relationships between operation and maintenance entities as edges; wherein the operation and maintenance entities include power equipment information, fault information and maintenance methods, and a probability value is assigned to each edge, and the probability value serves as the confidence level of the connection relationship between the operation and maintenance entities.

[0054] Based on each extracted operation and maintenance entity, the corresponding associated information is queried in the above-mentioned power operation and maintenance map. The associated information includes the entity itself and description information. The queried entity and description information pairs are spliced ​​in a dictionary format, for example, {"entity name":"description information"}. The spliced ​​text is stored as the context information of the current question, which will be used to fill in the predefined prompts later.

[0055] Further, in step S2, the association information is added to the predefined Chain-of-Thought prompt template The expression in can be expressed as:

[0056]

[0057] Among them, t+1 represents the number of rounds of the next round of answering process of the agent; p I is the question-answer instruction, p E For the sample answer, p Q For questions and options, pK represents the associated information, represents the updated reasoning process to be used in the t+1th round of answering, Represents the updated reflection text that will be used in the t+1 round of answering.

[0058] The agent receives the Chain-of-Thought prompt filled in the above steps, and generates the reasoning process and answer a for the t+1th answer in sequence; the format of answer a is an uppercase option string, such as 'ABD'. The reasoning process and answer a are stored in memory 1.

[0059] In this embodiment, in step S3, if the number of times the current question is answered does not exceed the set answer limit, the answer a generated according to the Chain-of-Thought prompt template for the t+1th answer is scored, and the calculation expression for the evaluation score score of the answer a is obtained as follows:

[0060]

[0061] Among them, a * is the candidate answer to the corresponding question in the question-answer pair, len(x) represents the function used to obtain the length of string x, and the value range of score is between [0,1].

[0062] Specifically, the method of performing reflective analysis based on the evaluation score to obtain the final answer to the question is as follows:

[0063] In this embodiment, the first threshold is set to 1 to determine whether the evaluation score reaches the set first threshold:

[0064] If the evaluation score score does not reach the set first threshold, it is necessary to generate a reflection text and conduct a new round of question and answer, that is, reflect and analyze the answer, generate a reflection text, update the chain-of-thought prompt template according to the reasoning process and the reflection text, obtain an updated Chain-of-Thought prompt template, return to step S2, and generate the reasoning process and answer for the next question according to the updated Chain-of-Thought prompt template;

[0065] The reflection text includes three parts: the problem of generating the answer, the cause of the problem and the corresponding solution. The problem of generating the answer includes two types: missing selection and wrong selection. According to the problem of generating the answer, the current problem is analyzed, the cause of the problem is obtained, and the corresponding solution is formulated;

[0066] If the evaluation score is 1, it means that answer a is the same length as answer a. * If it matches perfectly, the agent can stop answering, which is the final answer to the question.

[0067] In this embodiment, the expression for updating the Chain-of-Thought prompt template according to the reasoning process and the reflection text is as follows:

[0068]

[0069] in, represents the updated reasoning process to be used in the t+1th round of answering, represents the reasoning process generated during the t-th round of answering; represents the updated reflection text to be used in the t+1 round of answers, Represents the reflection text generated during the tth round of answers.

[0070] The proposed method was applied to a question-answering dataset in the field of operation and maintenance, and a performance test experiment was conducted. The ratio of completely correct answers to partially correct answers was systematically analyzed. The question-answering dataset consists of 1,844 high-quality question-answer pairs, which are subdivided into 994 single-choice questions and 850 multiple-choice questions, aiming to comprehensively evaluate the performance of the method. The experimental design covers 5 rounds of question-answering interaction to deeply explore the performance changes of the method under different iteration numbers.

[0071] like Figure 3 and Figure 4 As shown in the figure, the experimental results show that the method of the present invention has shown significant results when only Reflexion is used, with the highest proportion of completely correct answers reaching 55.80%. After 5 rounds of reflection iterations, the performance improvement has reached 25.97%. At the same time, in terms of missed selections, the proportion has been effectively reduced to 17.52%. Although the reduction in this proportion is relatively small, only 2%, after 5 rounds of reflection, it still reflects the trend of continuous improvement.

[0072] Furthermore, when Reflexion is combined with the knowledge graph (KG), the performance of the method of the present invention is significantly improved. Under this configuration, the proportion of completely correct answers climbed to 64.05%, and after 5 rounds of reflection iterations, the performance improvement was as high as 33.84%. It is particularly worth mentioning that the proportion of missed selections was significantly reduced to 6.45%, and during the 5 rounds of reflection, the reduction effect of this proportion reached 17.03%, demonstrating the superiority and stability of the method in dealing with complex multiple-choice questions.

[0073] In summary, the above experimental results fully verify the powerful ability of the method of the present invention in the application of knowledge in the field of power operation and maintenance. It not only effectively improves the accuracy of the answer, but also continuously optimizes the performance through the iterative reflection mechanism, providing strong support for the development and application of intelligent operation and maintenance systems.

[0074] Embodiment 2

[0075] Based on the same inventive concept as that of the first embodiment, the present invention also provides a power operation and maintenance knowledge capability measurement system based on Reflexion and knowledge graph, which is used to implement the steps of the power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph described in the first embodiment. Figure 5 As shown, the system includes the following modules:

[0076] The operation and maintenance entity identification and associated information query module 100 is used to identify and extract the operation and maintenance entity in any question-answer pair, and query the associated information corresponding to the operation and maintenance entity in the power operation and maintenance map according to the operation and maintenance entity;

[0077] The chain thinking prompt template construction and answer generation module 200 is used to add the associated information to the predefined chain thinking prompt template, and generate the reasoning process and answer for the next question according to the chain thinking prompt template;

[0078] The answer scoring and final answer determination module 300 is used to score the answer if the number of current answers to the question does not exceed the set answer limit, obtain the evaluation score of the answer, and perform reflection analysis based on the evaluation score to obtain the final answer to the question.

[0079] The present embodiment proposes a power operation and maintenance knowledge capability measurement system based on Reflexion and knowledge graph, which is used to implement the aforementioned power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph. Therefore, the specific implementation method of the power operation and maintenance knowledge capability measurement system based on Reflexion and knowledge graph can be seen from the aforementioned embodiment part of the power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph. For example, the operation and maintenance entity recognition and associated information query module 100, the chain thinking prompt template construction and answer generation module 200 and the answer scoring and final answer determination module 300 are respectively used to correspond to the implementation of steps S1, S2 and S3 in the power operation and maintenance knowledge capability measurement method based on Reflexion and knowledge graph in embodiment one. Therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part. In order to avoid redundancy, it will not be repeated here.

[0080] Embodiment 3

[0081] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps of the method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph described in Example 1 are implemented.

[0082] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0083] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0084] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0085] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0086] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A method for measuring knowledge capability of power operation and maintenance based on Reflexion and knowledge graph, characterized in that: The following steps are involved: S1: Identify and extract the operation and maintenance entity in any question-answer pair, and query the corresponding associated information in the power operation and maintenance map according to the operation and maintenance entity; S2: adding the associated information to a predefined chain thinking prompt template, and generating a reasoning process and an answer for the next question according to the chain thinking prompt template; S3: Determine whether the number of times the current question is answered exceeds the set answer limit: If not, score the answer to obtain an evaluation score for the answer, and reflect and analyze based on the evaluation score to obtain a final answer to the question; If so, the answer is the final answer to the question.

2. The method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph according to claim 1 is characterized in that: In S2, the associated information is added to a predefined chain thinking prompt template The expression in can be expressed as: Among them, t+1 represents the number of rounds of the next round of answering process of the agent; p I is the question-answer instruction, p E For the sample answer, p Q For questions and options, p K represents the associated information, represents the updated reasoning process to be used in the t+1th round of answering, Represents the updated reflection text that will be used in the t+1 round of answering.

3. The method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph according to claim 1 is characterized in that: In S3, the calculation expression for obtaining the evaluation score of the answer is as follows: Where a is the answer generated according to the chain thinking prompt template, a * is the candidate answer in the question-answer pair, len(x) represents the function used to obtain the length of string x, and the value range of score is between [0,1].

4. The method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph according to claim 1 is characterized in that: In S3, the method of performing reflective analysis based on the evaluation score to obtain the final answer to the question is as follows: Determine whether the evaluation score reaches a set first threshold: If not, reflectively analyze the answer to generate a reflection text, update the chain thinking prompt template according to the reasoning process and the reflection text, obtain an updated chain thinking prompt template, return to step S2, and generate the reasoning process and answer for the next question according to the updated chain thinking prompt template; If so, it means that the answer completely matches the answer in the question-answer pair, which is the final answer to the question.

5. The method for measuring electric power operation and maintenance knowledge capability based on Reflexion and knowledge graph according to claim 4 is characterized in that: The expression for updating the chained thinking prompt template according to the reasoning process and the reflection text is as follows: in, represents the updated reasoning process to be used in the t+1th round of answering, represents the reasoning process generated during the t-th round of answering; represents the updated reflection text to be used in the t+1 round of answers, Represents the reflection text generated during the tth round of answers.

6. The method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph according to claim 4 is characterized in that: The reflection text includes three parts: problems generated by the answer, causes of the problems and corresponding solutions. Among them, the problems of the answer include missing selections and wrong selections. According to the problems of the answer, the current problem is analyzed, the causes of the problems are obtained, and corresponding solutions are formulated.

7. The method for measuring electric power operation and maintenance knowledge capability based on Reflexion and knowledge graph according to claim 1 is characterized in that: The power operation and maintenance graph is a directed graph with operation and maintenance entities as nodes and the relationships between operation and maintenance entities as edges; wherein the operation and maintenance entities include power equipment information, fault information and maintenance methods, and a probability value is assigned to each edge, and the probability value serves as the confidence level of the connection relationship between the operation and maintenance entities.

8. The method for measuring electric power operation and maintenance knowledge capability based on Reflexion and knowledge graph according to claim 1 is characterized in that: The question-answer pair is a triple data structure consisting of a question, an option set, and an answer set.<Q,O,A> ; Where Q represents the power operation and maintenance problem raised by the user, O = {O1, O2, ..., O n } represents a set of options, n≥1, each option O i is a candidate answer to question Q, Represents the answer set, which is a collection of one or more options in option set O that are determined to be the correct answer.

9. A power operation and maintenance knowledge capability measurement system based on Reflexion and knowledge graph, characterized in that: The system is used to implement the method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph as described in any one of claims 1 to 8, wherein the system comprises the following modules: The operation and maintenance entity identification and associated information query module is used to identify and extract the operation and maintenance entity in any question-answer pair, and query the associated information corresponding to the operation and maintenance entity in the power operation and maintenance map according to the operation and maintenance entity; A chain thinking prompt template construction and answer generation module, used to add the associated information to a predefined chain thinking prompt template, and generate a reasoning process and answer for the next question according to the chain thinking prompt template; The answer scoring and final answer determination module is used to score the answer if the number of current answers to the question does not exceed the set answer limit, obtain the evaluation score of the answer, and reflect and analyze based on the evaluation score to obtain the final answer to the question.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method for measuring the knowledge capability of power operation and maintenance based on Reflexion and knowledge graph are implemented as described in any one of claims 1 to 8.

Images