Intelligent analysis method and system for electric locomotive maintenance practical training based on AI model

Abstract

The application relates to the field of intelligent management analysis, and proposes an electric locomotive maintenance practical training intelligent analysis method and system based on an AI model, which constructs an electric locomotive virtual practical training scene, avoids the problem that a large amount of manpower and material resources needs to be consumed during actual scene training, constructs an electric locomotive maintenance knowledge graph, generates dynamic operation risks, makes the practical training operation results and fault problems more consistent with the actual scene, avoids the problem that a single fault case cannot restore the complex problems and complex operations of the actual maintenance scene, is based on a double-base interaction mechanism, ensures the accuracy and reliability of the guidance information, further improves the accuracy of the guidance information through deep expansion, gives deeper information, avoids the problem that the depth of the guidance information in the field knowledge is insufficient, and objectively authenticates the practical training effect through automatic quantitative evaluation, and the application improves the practical training effect and the practical training quality of operation and maintenance personnel.

Description

Technical Field

[0001] This invention relates to the field of intelligent management and analysis, and in particular to an intelligent analysis method and system for electric locomotive maintenance training based on an AI model. Background Technology

[0002] With the rapid development of the rail transit industry, the number of electric locomotives is increasing day by day, and the corresponding safety requirements for electric locomotives are also gradually increasing. In order to ensure the safety of electric locomotive operation, one of the primary ways is to carry out maintenance work. Therefore, it is essential to increase the number and skill level of maintenance personnel, and maintenance training has naturally become a crucial link.

[0003] In existing technologies, the maintenance training of electric locomotive operation and maintenance personnel often adopts the traditional training method of on-site demonstration and training. This traditional training method requires a lot of manpower and material resources. First, it occupies the existing locomotives for a long time. The existing vehicles need to be in operation and cannot meet the maintenance training requirements of operation and maintenance personnel at all times. Second, the space in a single area of ​​the locomotive is limited, which cannot meet the needs of multiple people to carry out maintenance training at the same time. In addition, the existing training and teaching methods are still mainly based on text, images and oral explanations, and the fault cases are simple and cannot reproduce the complex problems and complex operations of the actual maintenance scenario, resulting in poor training effect.

[0004] Therefore, it is necessary to design an intelligent analysis method for electric locomotive maintenance training to improve the training effect and the training quality of maintenance personnel. Summary of the Invention

[0005] Based on this, the present invention proposes an intelligent analysis method and system for electric locomotive maintenance training based on an AI model. By constructing a virtual training scenario for electric locomotives, it provides each trainee with individual and comprehensive immersive training, avoiding the problem of consuming a large amount of manpower and material resources required for training in actual scenarios. Furthermore, by constructing an electric locomotive maintenance knowledge graph, it dynamically generates operational risks, making the training operation results and fault problems more consistent with actual scenarios. This avoids the problem of single fault cases failing to reproduce the complex problems and operations of actual maintenance scenarios. Based on a dual-base interaction mechanism, it generates guidance information through a large language model and the electric locomotive maintenance knowledge graph. This not only ensures the accuracy and reliability of the guidance information, but also further improves the precision of the guidance information through deep expansion, while providing deeper information and avoiding the problem of insufficient depth of guidance information in domain knowledge. Moreover, through automatic quantitative evaluation, it objectively certifies the training effect. The present invention improves the training effect and the training quality of maintenance personnel.

[0006] This invention proposes an intelligent analysis method for electric locomotive maintenance training based on an AI model, comprising:

[0007] A virtual training scenario for electric locomotives is constructed, and the operational behavior data of trainees are collected in real time. The virtual training scenario for electric locomotives is based on the smallest training unit, which is a single maintenance process operation node.

[0008] A knowledge graph for electric locomotive maintenance is constructed. Dynamic operational risks are generated based on the operational behavior data and the knowledge graph for electric locomotive maintenance. The dynamic operational risks are dynamically generated based on a decision model. The knowledge graph for electric locomotive maintenance is constructed based on the entity relationships of electric locomotive maintenance documents.

[0009] Guidance information is generated based on the large language model and the electric locomotive maintenance knowledge graph, with the large language model and the electric locomotive maintenance knowledge graph based on a dual-base interaction mechanism.

[0010] A quantitative assessment is conducted based on the dynamic operational risks and the guidance information to obtain a final training analysis report.

[0011] In summary, based on the aforementioned AI-based intelligent analysis method for electric locomotive maintenance training, this invention constructs a virtual training scenario for electric locomotives, providing each trainee with individual and comprehensive immersive training. This avoids the significant manpower and material costs associated with actual scenario training. Furthermore, by building an electric locomotive maintenance knowledge graph and dynamically generating operational risks, the training results and fault problems are more consistent with real-world scenarios, avoiding the problem of limited fault cases and inability to recreate complex issues and operations in actual maintenance scenarios. Based on a dual-base interaction mechanism, guidance information is generated through a large language model and the electric locomotive maintenance knowledge graph. This not only ensures the accuracy and reliability of the guidance information but also improves its precision through in-depth expansion, providing deeper information and avoiding the problem of insufficient depth in domain knowledge. Finally, through automatic quantitative evaluation, the training effect is objectively certified. This invention improves the training effect and the training quality for maintenance personnel. Specifically, a virtual training scenario for electric locomotives is constructed, and operational behavior data of trainees is collected in real time. This virtual training scenario is based on minimum training units, each representing a single maintenance process node. This provides each trainee with individual and comprehensive immersive training, avoiding the significant manpower and material costs associated with real-world training scenarios. A knowledge graph for electric locomotive maintenance is constructed, and dynamic operational risks are generated based on the operational behavior data and the knowledge graph. These dynamic operational risks are dynamically generated based on a decision-making model. The knowledge graph is constructed based on the entity relationships within electric locomotive maintenance documents, ensuring that the training results align with the actual operational process. The fault problems are more in line with real-world scenarios, avoiding the problem of single fault cases failing to recreate complex problems and operations in actual maintenance scenarios. Guidance information is generated based on the large language model and the electric locomotive maintenance knowledge graph. The large language model and the electric locomotive maintenance knowledge graph are based on a dual-base interaction mechanism, which not only ensures the accuracy and reliability of the guidance information, but also improves the precision of the guidance information through deep expansion, while providing deeper information and avoiding the problem of insufficient depth of guidance information in domain knowledge. The dynamic operation risk is quantitatively evaluated based on the guidance information to obtain the final training analysis report. This invention improves the training effect and the training quality of operation and maintenance personnel.

[0012] Furthermore, the step of constructing a virtual training scenario for electric locomotives and collecting real-time operational behavior data of trainees specifically includes:

[0013] A virtual model of the electric locomotive is constructed based on the shape of its components and their connection relationships. A virtual maintenance environment is then constructed based on the locomotive maintenance depot to generate a virtual training scenario for the electric locomotive.

[0014] Traverse the maintenance process relationship diagram, break down the maintenance process operation flow in the maintenance process relationship diagram into nodes, set a single maintenance process operation flow node as the minimum training unit, and connect each minimum training unit with the electric locomotive virtual model in an operation relationship. Generate a maintenance process prefabricated body by connecting the minimum training unit with the electric locomotive virtual model according to the maintenance process operation flow.

[0015] Furthermore, the steps for constructing the electric locomotive maintenance knowledge graph specifically include:

[0016] Extract a single text sentence from the electric locomotive maintenance document, convert the single text sentence into a character vector, perform deep semantic encoding on the single text sentence according to the BERT model to obtain a category vector, and label the single text sentence according to the category vector to obtain an entity;

[0017] Entity unification is performed based on multi-level similarity to obtain entity relationships. A knowledge graph of electric locomotive maintenance is generated based on the entities and entity relationships. The multi-level similarity includes string similarity, statistical feature similarity, and deep semantic vector similarity. The string similarity is used to remove explicit synonyms, the statistical feature similarity is used to remove near-synonyms, and the deep semantic vector similarity is used to match vector representations.

[0018] Furthermore, the specific data processing steps for the electric locomotive maintenance knowledge graph include:

[0019] The input electric locomotive maintenance document is parsed by the unstructured data loader, and its content is read and converted into structured text data. The long text in the structured text data is segmented by the text segmenter to obtain text blocks. The text blocks are converted into high-dimensional vector representations by the vectorization module and stored in a vector library. The vector library generates matching index keys based on multi-level similarity, and each high-dimensional vector representation has a unique matching index key.

[0020] Furthermore, the step of generating dynamic operational risks based on the operational behavior data and the electric locomotive maintenance knowledge graph specifically includes:

[0021] A reinforcement learning state space is constructed based on operational behavior data, and an optimized reward function is constructed based on safety rules in the electric locomotive maintenance knowledge graph to optimize the decision-making model;

[0022] Based on the decision-making model, risk iteration is performed in a virtual training scenario for electric locomotives to obtain dynamic operational risks.

[0023] Furthermore, the step of generating guidance information based on the large language model and the electric locomotive maintenance knowledge graph specifically includes:

[0024] Obtain fault information corresponding to dynamic operational risks or fault problems actively input by trainees;

[0025] The fault information and fault problems are retrieved through a multi-level search based on the electric locomotive maintenance knowledge graph to obtain basic guidance knowledge. The multi-level search is used to filter the first-level guidance knowledge and the second-level guidance knowledge related to the fault information and fault problems. The second-level guidance knowledge is the secondary related content of the first-level guidance knowledge.

[0026] The basic guidance knowledge is input into a large language model, which then expands the knowledge boundary to generate in-depth guidance knowledge.

[0027] The basic guidance knowledge and the deep guidance knowledge are fused and generated according to the reordering model to obtain guidance information.

[0028] Furthermore, the step of quantitatively assessing the dynamic operational risks and the guidance information specifically includes:

[0029] A quantitative assessment is conducted based on dynamic operational risks and guidance information to obtain a practical training evaluation score.

[0030] The parameters of the optimization reward function of the decision-making model are adjusted based on the training evaluation scores.

[0031] A final training analysis report is generated based on the training evaluation score. Operation examples are generated based on the final training analysis report. The operation examples are converted into graph increments and the electric locomotive maintenance knowledge graph is updated.

[0032] The final training analysis report is transmitted to the training assessment and evaluation management system according to the standardized interface.

[0033] An AI-based intelligent analysis system for electric locomotive maintenance training includes:

[0034] The virtual training module is used to construct a virtual training scenario for electric locomotives and collect operational behavior data of trainees in real time. The virtual training scenario for electric locomotives is based on a minimum training unit, which is a single maintenance process operation node.

[0035] The dynamic operation risk generation module is used to construct a knowledge graph for electric locomotive maintenance. It generates dynamic operation risks based on the operation behavior data and the knowledge graph for electric locomotive maintenance. The dynamic operation risks are dynamically generated based on a decision model, and the knowledge graph for electric locomotive maintenance is constructed based on the entity relationships of electric locomotive maintenance documents.

[0036] The guidance information generation module is used to generate guidance information based on the large language model and the electric locomotive maintenance knowledge graph. The large language model and the electric locomotive maintenance knowledge graph are based on a dual-base interaction mechanism.

[0037] The quantitative assessment module is used to perform quantitative assessments based on the dynamic operational risks and the guidance information to obtain a final training analysis report.

[0038] The present invention also provides a storage medium that stores one or more programs, which, when executed by a processor, implement the AI ​​model-based intelligent analysis method for electric locomotive maintenance training as described above.

[0039] The present invention also provides a computer device, the computer device including a memory and a processor, wherein:

[0040] The memory is used to store computer programs;

[0041] When the processor executes the computer program stored in the memory, it implements the AI-based intelligent analysis method for electric locomotive maintenance training as described above. Attached Figure Description

[0042] Figure 1 The flowchart shows the intelligent analysis method for electric locomotive maintenance training based on an AI model proposed in the first embodiment of the present invention.

[0043] Figure 2 This is a structural diagram of the intelligent analysis system for electric locomotive maintenance training based on an AI model, as proposed in the second embodiment of the present invention.

[0044] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0045] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0046] It should be noted that when a component is said to be "fixed to" another component, it can be directly on the other component or there may be an intervening component. When a component is said to be "connected to" another component, it can be directly connected to the other component or there may be an intervening component. The terms "vertical," "horizontal," "left," "right," and similar expressions used in this document are for illustrative purposes only.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0048] Please see Figure 1 The diagram shows a flowchart of the intelligent analysis method for electric locomotive maintenance training based on an AI model proposed in the first embodiment of the present invention. This intelligent analysis method for electric locomotive maintenance training based on an AI model includes steps S01 to S04, wherein:

[0049] Step S01: Construct a virtual training scenario for electric locomotives and collect real-time operational behavior data of trainees;

[0050] It should be noted that in this embodiment, the electric locomotive virtual training scenario is based on the smallest training unit, which is a single maintenance process operation node. A virtual model of the electric locomotive is constructed based on the shape of the locomotive's parts and the connection relationship between the parts. Then, a virtual maintenance environment is constructed based on the locomotive maintenance library to generate the electric locomotive virtual training scenario.

[0051] Traverse the maintenance process relationship diagram, break down the maintenance process operation flow in the maintenance process relationship diagram into nodes, set a single maintenance process operation flow node as the minimum training unit, and connect each minimum training unit with the electric locomotive virtual model in an operation relationship. Generate a maintenance process prefabricated body by connecting the minimum training unit with the electric locomotive virtual model according to the maintenance process operation flow.

[0052] Step S02: Construct a knowledge graph for electric locomotive maintenance, and generate dynamic operational risks based on operational behavior data and the knowledge graph for electric locomotive maintenance;

[0053] It should be noted that in this embodiment, the dynamic operation risk is dynamically generated based on the decision model, and the electric locomotive maintenance knowledge graph is constructed based on the entity relationship of the electric locomotive maintenance document. A single text sentence is extracted from the electric locomotive maintenance document and converted into a character vector. The single text sentence is then subjected to deep semantic encoding according to the BERT model to obtain a category vector. The single text sentence is then labeled according to the category vector to obtain an entity.

[0054] Entity unification is performed based on multi-level similarity to obtain entity relationships. A knowledge graph of electric locomotive maintenance is generated based on the entities and entity relationships. The multi-level similarity includes string similarity, statistical feature similarity, and deep semantic vector similarity. The string similarity is used to remove explicit synonyms, the statistical feature similarity is used to remove near-synonyms, and the deep semantic vector similarity is used to match vector representations.

[0055] The input electric locomotive maintenance document is parsed by the unstructured data loader, and its content is read and converted into structured text data. The long text in the structured text data is segmented by the text segmenter to obtain text blocks. The text blocks are converted into high-dimensional vector representations by the vectorization module and stored in a vector library. The vector library generates matching index keys based on multi-level similarity, and each high-dimensional vector representation has a unique matching index key.

[0056] A reinforcement learning state space is constructed based on operational behavior data, and an optimized reward function is constructed based on safety rules in the electric locomotive maintenance knowledge graph to optimize the decision-making model;

[0057] Based on the decision-making model, risk iteration is performed in a virtual training scenario for electric locomotives to obtain dynamic operational risks.

[0058] Step S03: Generate guidance information based on the large language model and the electric locomotive maintenance knowledge graph;

[0059] It should be noted that in this embodiment, the large language model and the electric locomotive maintenance knowledge graph are based on a dual-base interaction mechanism to obtain fault information corresponding to dynamic operational risks or fault problems actively input by trainees.

[0060] The fault information and fault problems are retrieved through a multi-level search based on the electric locomotive maintenance knowledge graph to obtain basic guidance knowledge. The multi-level search is used to filter the first-level guidance knowledge and the second-level guidance knowledge related to the fault information and fault problems. The second-level guidance knowledge is the secondary related content of the first-level guidance knowledge.

[0061] The basic guidance knowledge is input into a large language model, which then expands the knowledge boundary to generate in-depth guidance knowledge.

[0062] The basic guidance knowledge and the deep guidance knowledge are fused and generated according to the reordering model to obtain guidance information.

[0063] Step S04: Conduct a quantitative assessment based on dynamic operational risks and guidance information to obtain the final training analysis report;

[0064] It should be noted that in this embodiment, a quantitative assessment is performed based on dynamic operational risks and guidance information to obtain a practical training assessment score.

[0065] The parameters of the optimization reward function of the decision-making model are adjusted based on the training evaluation scores.

[0066] A final training analysis report is generated based on the training evaluation score. Operation examples are generated based on the final training analysis report. The operation examples are converted into graph increments and the electric locomotive maintenance knowledge graph is updated.

[0067] The final training analysis report is transmitted to the training assessment and evaluation management system according to the standardized interface.

[0068] In summary, based on the aforementioned AI-based intelligent analysis method for electric locomotive maintenance training, this invention constructs a virtual training scenario for electric locomotives, providing each trainee with individual and comprehensive immersive training. This avoids the significant manpower and material costs associated with actual scenario training. Furthermore, by building an electric locomotive maintenance knowledge graph and dynamically generating operational risks, the training results and fault problems are more consistent with real-world scenarios, avoiding the problem of limited fault cases and inability to recreate complex issues and operations in actual maintenance scenarios. Based on a dual-base interaction mechanism, guidance information is generated through a large language model and the electric locomotive maintenance knowledge graph. This not only ensures the accuracy and reliability of the guidance information but also improves its precision through in-depth expansion, providing deeper information and avoiding the problem of insufficient depth in domain knowledge. Finally, through automatic quantitative evaluation, the training effect is objectively certified. This invention improves the training effect and the training quality for maintenance personnel. Specifically, a virtual training scenario for electric locomotives is constructed, and operational behavior data of trainees is collected in real time. This virtual training scenario is based on minimum training units, each representing a single maintenance process node. This provides each trainee with individual and comprehensive immersive training, avoiding the significant manpower and material costs associated with real-world training scenarios. A knowledge graph for electric locomotive maintenance is constructed, and dynamic operational risks are generated based on the operational behavior data and the knowledge graph. These dynamic operational risks are dynamically generated based on a decision-making model. The knowledge graph is constructed based on the entity relationships within electric locomotive maintenance documents, ensuring that the training results align with the actual operational process. The fault problems are more in line with real-world scenarios, avoiding the problem of single fault cases failing to recreate complex problems and operations in actual maintenance scenarios. Guidance information is generated based on the large language model and the electric locomotive maintenance knowledge graph. The large language model and the electric locomotive maintenance knowledge graph are based on a dual-base interaction mechanism, which not only ensures the accuracy and reliability of the guidance information, but also improves the precision of the guidance information through deep expansion, while providing deeper information and avoiding the problem of insufficient depth of guidance information in domain knowledge. The dynamic operation risk is quantitatively evaluated based on the guidance information to obtain the final training analysis report. This invention improves the training effect and the training quality of operation and maintenance personnel.

[0069] Please see Figure 2 The diagram shows a schematic representation of the intelligent analysis system for electric locomotive maintenance training based on an AI model, as proposed in the second embodiment of the present invention. The system includes:

[0070] Virtual training module 10 is used to construct a virtual training scenario for electric locomotives and collect operational behavior data of trainees in real time. The virtual training scenario for electric locomotives is based on a minimum training unit, which is a single maintenance process operation node.

[0071] The dynamic operation risk generation module 20 is used to construct a knowledge graph for electric locomotive maintenance, and generate dynamic operation risks based on the operation behavior data and the knowledge graph for electric locomotive maintenance. The dynamic operation risks are dynamically generated based on a decision model, and the knowledge graph for electric locomotive maintenance is constructed based on the entity relationships of electric locomotive maintenance documents.

[0072] The guidance information generation module 30 is used to generate guidance information based on the large language model and the electric locomotive maintenance knowledge graph, wherein the large language model and the electric locomotive maintenance knowledge graph are based on a dual-base interaction mechanism.

[0073] The quantitative assessment module 40 is used to perform a quantitative assessment based on the dynamic operational risks and the guidance information to obtain a final training analysis report.

[0074] The present invention also proposes a computer storage medium storing one or more programs, which, when executed by a processor, implement the above-mentioned intelligent analysis method for electric locomotive maintenance training based on an AI model.

[0075] The present invention also proposes a computer device, including a memory and a processor, wherein the memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to realize the above-mentioned intelligent analysis method for electric locomotive maintenance training based on AI models.

[0076] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain stored, communicated, propagated, or transmitted programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0077] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0078] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0079] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0080] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. An intelligent analysis method for electric locomotive maintenance training based on an AI model, characterized in that, include: A virtual training scenario for electric locomotives is constructed, and the operational behavior data of trainees are collected in real time. The virtual training scenario for electric locomotives is based on the smallest training unit, which is a single maintenance process operation node. The steps of constructing a virtual training scenario for electric locomotives and collecting real-time operational behavior data of trainees specifically include: A virtual model of the electric locomotive is constructed based on the shape of its components and their connection relationships. A virtual maintenance environment is then constructed based on the locomotive maintenance depot to generate a virtual training scenario for the electric locomotive. Traverse the maintenance process relationship diagram, break down the maintenance process operation flow in the maintenance process relationship diagram into nodes, set a single maintenance process operation flow node as the minimum training unit, and connect each minimum training unit with the electric locomotive virtual model in an operation relationship. Generate a maintenance process prefabricated body based on the maintenance process operation flow and the minimum training unit and the electric locomotive virtual model. A knowledge graph for electric locomotive maintenance is constructed. Dynamic operational risks are generated based on the operational behavior data and the knowledge graph for electric locomotive maintenance. The dynamic operational risks are dynamically generated based on a decision model. The knowledge graph for electric locomotive maintenance is constructed based on the entity relationships of electric locomotive maintenance documents. The step of generating dynamic operational risks based on the operational behavior data and the electric locomotive maintenance knowledge graph specifically includes: A reinforcement learning state space is constructed based on operational behavior data, and an optimized reward function is constructed based on safety rules in the electric locomotive maintenance knowledge graph to optimize the decision-making model; Based on the aforementioned decision-making model, risk iteration is performed in a virtual training scenario for electric locomotives to obtain dynamic operational risks. Guidance information is generated based on the large language model and the electric locomotive maintenance knowledge graph, with the large language model and the electric locomotive maintenance knowledge graph based on a dual-base interaction mechanism. The step of generating guidance information based on the large language model and the electric locomotive maintenance knowledge graph specifically includes: Obtain fault information corresponding to dynamic operational risks or fault problems actively input by trainees; The fault information and fault problems are retrieved through a multi-level search based on the electric locomotive maintenance knowledge graph to obtain basic guidance knowledge. The multi-level search is used to filter the first-level guidance knowledge and the second-level guidance knowledge related to the fault information and fault problems. The second-level guidance knowledge is the secondary related content of the first-level guidance knowledge. The basic guidance knowledge is input into a large language model, which then expands the knowledge boundary to generate in-depth guidance knowledge. The basic guidance knowledge and the deep guidance knowledge are fused and generated according to the reordering model to obtain guidance information; A quantitative assessment is conducted based on the dynamic operational risks and the guidance information to obtain a final training analysis report. The step of quantitatively assessing the dynamic operational risks and the guidance information specifically includes: A quantitative assessment is conducted based on dynamic operational risks and guidance information to obtain a practical training evaluation score. The parameters of the optimization reward function of the decision-making model are adjusted based on the training evaluation scores. A final training analysis report is generated based on the training evaluation score. Operation examples are generated based on the final training analysis report. The operation examples are converted into graph increments and the electric locomotive maintenance knowledge graph is updated. The final training analysis report is transmitted to the training assessment and evaluation management system according to the standardized interface.

2. The intelligent analysis method for electric locomotive maintenance training based on an AI model as described in claim 1, characterized in that, The steps for constructing the knowledge graph of electric locomotive maintenance specifically include: Extract a single text sentence from the electric locomotive maintenance document, convert the single text sentence into a character vector, perform deep semantic encoding on the single text sentence according to the BERT model to obtain a category vector, and label the single text sentence according to the category vector to obtain an entity; Entity unification is performed based on multi-level similarity to obtain entity relationships. A knowledge graph of electric locomotive maintenance is generated based on the entities and entity relationships. The multi-level similarity includes string similarity, statistical feature similarity, and deep semantic vector similarity. The string similarity is used to remove explicit synonyms, the statistical feature similarity is used to remove near-synonyms, and the deep semantic vector similarity is used to match vector representations.

3. The intelligent analysis method for electric locomotive maintenance training based on an AI model as described in claim 2, characterized in that, The specific data processing steps for the electric locomotive maintenance knowledge graph include: The input electric locomotive maintenance document is parsed by the unstructured data loader, and its content is read and converted into structured text data. The long text in the structured text data is segmented by the text segmenter to obtain text blocks. The text blocks are converted into high-dimensional vector representations by the vectorization module and stored in a vector library. The vector library generates matching index keys based on multi-level similarity, and each high-dimensional vector representation has a unique matching index key.

4. An intelligent analysis system for electric locomotive maintenance training based on an AI model, characterized in that, include: A virtual training scenario for electric locomotives is constructed, and the operational behavior data of trainees are collected in real time. The virtual training scenario for electric locomotives is based on the smallest training unit, which is a single maintenance process operation node. The steps of constructing a virtual training scenario for electric locomotives and collecting real-time operational behavior data of trainees specifically include: A virtual model of the electric locomotive is constructed based on the shape of its components and their connection relationships. A virtual maintenance environment is then constructed based on the locomotive maintenance depot to generate a virtual training scenario for the electric locomotive. Traverse the maintenance process relationship diagram, break down the maintenance process operation flow in the maintenance process relationship diagram into nodes, set a single maintenance process operation flow node as the minimum training unit, and connect each minimum training unit with the electric locomotive virtual model in an operation relationship. Generate a maintenance process prefabricated body based on the maintenance process operation flow and the minimum training unit and the electric locomotive virtual model. A knowledge graph for electric locomotive maintenance is constructed. Dynamic operational risks are generated based on the operational behavior data and the knowledge graph for electric locomotive maintenance. The dynamic operational risks are dynamically generated based on a decision model. The knowledge graph for electric locomotive maintenance is constructed based on the entity relationships of electric locomotive maintenance documents. The step of generating dynamic operational risks based on the operational behavior data and the electric locomotive maintenance knowledge graph specifically includes: A reinforcement learning state space is constructed based on operational behavior data, and an optimized reward function is constructed based on safety rules in the electric locomotive maintenance knowledge graph to optimize the decision-making model; Based on the aforementioned decision-making model, risk iteration is performed in a virtual training scenario for electric locomotives to obtain dynamic operational risks. Guidance information is generated based on the large language model and the electric locomotive maintenance knowledge graph, with the large language model and the electric locomotive maintenance knowledge graph based on a dual-base interaction mechanism. The step of generating guidance information based on the large language model and the electric locomotive maintenance knowledge graph specifically includes: Obtain fault information corresponding to dynamic operational risks or fault problems actively input by trainees; The fault information and fault problems are retrieved through a multi-level search based on the electric locomotive maintenance knowledge graph to obtain basic guidance knowledge. The multi-level search is used to filter the first-level guidance knowledge and the second-level guidance knowledge related to the fault information and fault problems. The second-level guidance knowledge is the secondary related content of the first-level guidance knowledge. The basic guidance knowledge is input into a large language model, which then expands the knowledge boundary to generate in-depth guidance knowledge. The basic guidance knowledge and the deep guidance knowledge are fused and generated according to the reordering model to obtain guidance information; A quantitative assessment is conducted based on the dynamic operational risks and the guidance information to obtain a final training analysis report. The step of quantitatively assessing the dynamic operational risks and the guidance information specifically includes: A quantitative assessment is conducted based on dynamic operational risks and guidance information to obtain a practical training evaluation score. The parameters of the optimization reward function of the decision-making model are adjusted based on the training evaluation scores. A final training analysis report is generated based on the training evaluation score. Operation examples are generated based on the final training analysis report. The operation examples are converted into graph increments and the electric locomotive maintenance knowledge graph is updated. The final training analysis report is transmitted to the training assessment and evaluation management system according to the standardized interface.

5. A storage medium, characterized in that, The storage medium stores one or more programs, which, when executed by a processor, implement the AI-based intelligent analysis method for electric locomotive maintenance training as described in any one of claims 1-3.

6. A computer device, characterized in that, The computer device includes a memory and a processor, wherein: The memory is used to store computer programs; When the processor executes the computer program stored in the memory, it implements the AI-based intelligent analysis method for electric locomotive maintenance training as described in any one of claims 1-3.

Images